Upcycling of steel slag to activate persulfate for the degradation of chlortetracycline hydrochloride: Performance and mechanism

In recent years, the environmental problems caused by antibiotics have attracted wide attention. Tetracycline (TCs) antibiotics (mainly including tetracycline (TC), chlortetracycline (CTC), oxytetracycline (OTC), etc.) are widely used in humans and animals around the world to prevent diseases or promote growth (Yang et al. 2021). A large amount of TCs will eventually flow into surface water through feces and urine, leading to the increase of antibiotic-resistant bacteria. Therefore, there is an urgent need to develop effective methods to remove antibiotics. Many methods have been used to remove TCs from wastewater, such as adsorption, biological, advanced oxidation processes (AOPs), etc. Although the adsorption method is simple and efficient, it cannot really remove antibiotics. However, AOPs can oxidize TC molecules into small molecules of organic matter, and can even oxidize TCs into non-toxic and harmless inorganic substances such as CO₂ and H₂O, thus achieving the removal of antibiotics in a real sense (Anjali & Shanthakumar 2019).

Sulfate radical (⁠⁠)-based AOPs have received widespread attention for the treatment of emerging pollutants such as persistent organic pollutants, endocrine disruptors (EDCs), antibiotics, and microplastics. Compared to the traditional hydroxyl radical (•OH), has higher redox potential (E⁰ = 2.60–3.10 V) and longer contact lifetime (t_1/2 = 30–40 μs) allowing excellent mass transfer and contact between radicals and the target pollutants, thus inducing its higher reactivity and selectivity. In addition, persulfate (PDS) as an oxidant has the advantages of lower costs of storage and transportation and less dependence of efficiency on the operational parameters. Commonly, persulfates can be activated by various methods including an input of energy in the form of ultraviolet, heat, or ultrasound and alkali and excessive metal ions (such as Fe²⁺, Cu²⁺, Co²⁺, etc.) to produce reactive oxygen species. Among them, the transition metal activation method has drawn significant attention due to its less complexity and more economical efficiency (Zhang et al. 2020a).

Steel slag is one of the main by-products of the steelmaking industry. On average, about 0.15–0.20 ton of steel slag is generated for every ton of crude steel produced, and more than 1.6 billion tons of steel slag are produced globally each year. At present, steel slag is mainly used for building materials, valuable metal recovery, environmental remediation, and so on. However, these treatment methods cannot help in the efficient utilization of steel slag, and most of the steel slag has to be stored in landfill (Gao et al. 2023). However, the landfill treatment of steel slag not only occupies a large amount of land resources, but also causes serious pollution to surface water, groundwater, and soil. Therefore, it is necessary to develop some new treatment methods to improve the comprehensive utilization rate of steel slag.

Metal catalysts for activating persulfates are generally prepared from pure chemical reagents, resulting in the high costs of catalysts. Various excessive metals (Fe, Co, Cu, Mn, etc.) are commonly found in industrial solid wastes, which can be used to prepare metal catalysts for activating persulfates, thus reducing the cost of wastewater treatment. For example, steel slag contains high content of iron, which can activate persulfate to produce free radicals with high activity to deoxidize and degrade stubborn organic pollutants (Zhang et al. 2023). However, there is also a large amount of calcium oxide in steel slag coated on the surface of iron compounds, inhibiting the potential catalytic activity of Fe. Hence, it is necessary to remove Ca impurities before using (Wang et al. 2021). Traditional strong acids such as hydrochloric acid and sulfuric acid can remove Ca from the surface of steel slag particles (Ragipani et al. 2020), but at the same time, there will be the loss of active metals such as Fe, leading to the decrease in catalytic activity. However, salicylic acid as a weak acid can selectively remove the calcium-containing compounds in steel slag. In this study, salicylic acid was used to modify steel slag powder (SSP) to obtain a heterogeneous catalyst ‘salicylic acid modified steel slag powder’ (SA-SSP) to effectively remove chloramphenicol hydrochloride from wastewater. X-ray diffraction (XRD), scanning electron microscopy (SEM), x-ray energy spectrometer (EDS), and Fourier transform infrared spectroscopy (FT-IR) results show that salicylic acid can selectively remove the calcium silicate mineral phase from SSP, and more iron and manganese metal catalytic active sites are exposed to efficiently degrade CTC from wastewater.

Steel slag used in this study was provided by a steel plant in the Yunnan Province of China. The chemical composition of the steel slag was determined by X-ray fluorescence (XRF PW 1400 spectrometer, Philips), and is listed in Table 1. The results show that the raw material of steel slag contains a large amount of calcium oxide and iron oxide, which accounts for about 60% of the weight of the steel slag. The following reagents are analytically pure and purchased from Aladdin, including persulfate (PDS, K₂S₂O₈), CTC (C₂₂H₂₄Cl₂N₂O₈), sodium thiosulfate (Na₂S₂O₃•5H₂O) methanol (MeOH), ethanol (EtOH), p-benzoquinone (BQ), furfuryl alcohol (FFA), tert-butanol (TBA), sodium carbonate (Na₂CO₃), sodium nitrate (NaNO₃), sodium dihydrogen phosphate (NaH₂PO₄), sodium sulfate (Na₂SO₄), and sodium chloride (NaCl).

Steel slag was ground in a ball mill for 3 h and passed through a 120 mesh sieve to obtain SSP raw material. Then, the obtained SSP was fully washed with ultrapure water at a solid–liquid ratio of 1:30 and dried in an oven at 105 °C for 12 h. The catalyst was prepared by adding 5 g dried SSP into 100 mL of 10–80 g/L salicylic acid solution, and then stirred for 0.5–7 h with a magnetic stirrer at 300 rpm at room temperature. The residue was washed with ethanol and ultrapure water for three times, and then dried at 105 °C for 12 h to obtain SA-SSP.

The crystal phase of the catalyst was determined by a Rigaku Ultima IV automatic X-ray diffractometer using a Cu Kα1 ray (incident wavelength 1.5406 A) at 10 − 80° and scanning speed 5°/min. The infrared spectrum of the sample was analyzed by the Fourier transform infrared spectrometer (Bruker tensor 27, German). The experiment was tested using an X-ray photoelectron spectrometer (Thermo Fisher Scientific K − Alpha). X-ray source: monochromatic Al Kα source (Mono Al Kα); Filament power: 72 W; full spectrum scanning: pass energy 100.0 eV, step size 1.0 eV; Narrow spectrum scanning: pass energy 30 eV, step size 0.1 eV. The binding energy was corrected by using surface contaminated carbon C1s (284.8 eV) as the standard, and the atomic percentage was calculated based on the results of a 400 × 400 μm region. The narrow spectrum of each element was fitted by XPS PEAK software to explore the valence distribution and material composition of each element. A Czech TESCAN MIRA LMS scanning electron microscope was used to characterize the surface morphology of the material, with a magnification of 10–20,000 times. X-ray energy dispersions (EDS) were used to determine the composition and content of chemical elements in the selected area through point analysis and plane analysis of the sample. Zetasizer Nano S (Malvern Instruments, UK) was used to measure the Zeta potential of the catalyst and determine the charge of the catalyst at different pH values. A Bruker A300 electron paramagnetic resonance spectrometer was used to examine the species of free radicals in the system using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as the trapping agent for •OH and ⁠, and 2,2,6,6-tetramethylpiperidine (TEMP) as the trapping agent for ¹O₂.

The degradation reaction was carried out in a conical flask by first weighing a specific amount of catalyst, then adding 100 mL of CTC solution (50 mg/L), followed by drawing a predetermined volume of PDS solution to initiate the degradation reaction. The temperature of the reaction system should be controlled by a water bath constant temperature oscillator (set temperature is 25 °C, speed is 150 rpm), and the pH of the system should be adjusted by 0.1 M sulfuric acid or sodium hydroxide. At a given reaction time, 2.0 mL of liquid was sucked out with a syringe with a 0.22 μm filter membrane, then transferred to a cuvette, and 0.1 mL of methanol was added to remove free radicals. The remaining concentration of CTC solution was immediately measured with a UV-visible photometer at a wavelength of 367 nm.

Factors such as the concentration of salicylic acid and acid-modification time may have a significant influence on the dissolution of calcium-containing components in steel slag, and thus affect the degradation performance of SA-SSP.

In addition, the influence of modification time on the residue mass of steel slag was also investigated. As shown in Figure 1(c), when the concentration of salicylic acid was set at 60 g/L, the mass of steel slag decreased with the extension of modification time. After 5 h of modification, the mass of SSP and the concentration of Ca²⁺ in filtrate no longer changed significantly, indicating that the calcium-containing component that can be dissolved by salicylic acid has basically disappeared. Similarly, the catalysts prepared for different modification times were used for the removal of CTC, as shown in Figure 1(d). It can be seen that when the time was extended from 0.5 to 5 h, the removal efficiency of CTC also gradually improved; but when it was extended to 7 h, the removal rate changed very little, indicating that the reaction between salicylic acid and steel slag has basically ended for 5 h. Therefore, the catalyst prepared at optimal conditions (salicylic acid concentration of 60 g/L and acid modification time of 5 h) was used for the following CTC degradation research.

The XRD patterns of steel slag before and after modification are shown in Supplementary material, Figure S1(a). The main mineral phases of SSP raw materials are larnite (Ca₂SiO₄), srebrodolskite (Ca₂Fe₂O₅), brownmillerite (Ca₂(Fe, Al)₂O₅), monticellite (CaMgSiO₄), ferric tetroxide (Fe₃O₄), ferrous oxide (FeO), and so on. After salicylic acid modification, the intensity of the characteristic peak of Ca₂SiO₄ in SA-SSP significantly decreased, while the intensity of the diffraction peaks of iron bearing phases (Fe₃O₄ and FeO) relatively enhanced (Xu et al. 2020). The results also proved the dissolution of calcium-containing components in SSP.

In order to further verify the dissolution of calcium-containing components in steel slag, the obtained SA-SSP was characterized by FT-IR, and the results are shown in Supplementary material, Figure S1(b). The absorption peaks at 3,440 and 1,637 cm⁻¹ are ascribed to the stretching and bending vibration of –OH. The band that appeared at 440 cm⁻¹ is attributed to the absorption peak of Si–O (Ma et al. 2022). The band at 517 cm⁻¹ is associated with Ca–O vibration, and it can be seen that the absorption peak of Ca–O is significantly weakened after modification, further confirming the decrease of calcium content in the steel slag. Both bands at 876 and 579 cm⁻¹ are ascribed to the vibration peaks of Fe–O, and the peak strength is slightly enhanced after modification. The absorption peak at 1,424 cm⁻¹ can be attributed to the metallic stretching of the Fe–Ca bond, and the enhancement of the band also proved the exposure of more Fe active sites. In addition, it can be found from Table 2 that the content of CaO in steel slag decreased from 36.33 to 20.54%, while the content of Fe₂O₃ increased from 21.72 to 35.15%. Therefore, combined with the characterization data of FT-IR and XRF, it is fully demonstrated that the Ca content of steel slag modified by salicylic acid decreases significantly while the Fe content increases relatively, which is also the reason for the improved removal rate of CTC from the modified steel slag (Xiao et al. 2020).

The surface morphology and microstructure of steel slag and SA-SSP are shown in Supplementary material, Figure S2. Compared with the flat and smooth surface of SSP (Supplementary material, Figure S2(a)), SA-SSP exhibited obviously porous and rough structure (Supplementary material, Figure S2(b)). EDS mapping analysis was carried out on the surface of steel slag particles. As shown in Supplementary material, Figure S2(c), the surface of unmodified steel slag particles is coated with a large amount of Ca (25.05%) and O (42.66%), while the contents of Fe and Mn elements are less (12.70 and 4.73%, respectively). After modification (Supplementary material, Figure S2(d)), the content of Ca on the steel slag surface decreased to 8.32%, while the relative contents of Fe and Mn increased significantly, and especially, the content of Fe increased from 12.70 to 65.63%. This indicates that the steel slag treated with salicylic acid can reduce the content of Ca element and expose more Fe and Mn metal active sites.

The influence of CTC concentration on the degradation effect was investigated, as shown in Figure 2(c). It can be seen that the degradation efficiency decreases as the initial concentration of CTC increases. When the initial concentration of CTC is below 50 mg/L, the removal rate can reach higher than 90%, indicating that the SA-SSP/PDS system is suitable for the removal of low concentration of CTC, and the concentration of antibiotics in the actual environment is mostly trace level (Javid et al. 2016).

The recyclability and easy recovery of catalysts are more conducive to practical applications and can endow with the low cost of antibiotic wastewater treatment (Yu et al. 2019). As shown in Figure 4(b), after four cycles, the removal rate of CTC remains at about 80%, indicating that SA-SSP has a better recycling ability. In addition, magnetite in SA-SSP also gives it magnetism, which can be determined by VSM, as shown in Figure 4(c). VSM results show that SA-SSP is sensitive to the magnetic field, and the saturation magnetization (Ms) can reach 7.36 emu/g, which indicates that SA-SSPs in aqueous solution can be separated and recovered by magnets. Therefore, SA-SSP is a kind of catalyst with certain recyclability and easy recovery and has the potential of practical applications.

After the addition of BQ, the degradation of CTC was weakly inhibited, indicating that accounted less in the system. After FFA was added to the system, it was also observed that the degradation of CTC was greatly inhibited (down to 49.38%), indicating that there was more non-free radical ¹O₂ in the system, whose contribution to the removal of CTC was 30.26%. The quenching experiments showed that •OH, ⁠, ⁠, and ¹O₂ existed in the SA-SSP/PDS system, and •OH and ¹O₂ played a major role in the degradation of CTC.

To further verify the existence of these different free radicals, electron spin resonance spectroscopy (EPR) was used to characterize them; 5,5-dimethyl-1-pyrroline n-oxide (DMPO) was selected as the trapping agent for •OH and ⁠, and TEMP was selected as the trapping agent for ¹O₂. The results are shown in Figure 5(b) and 5(c). When only PDS is added without SA-SSP as the catalyst, no obvious absorption peak appears, proving that •OH and are not produced. After adding SA-SSP (5 min) to the system, a peak-to-height ratio of 1:2:2:1 appears, which is the characteristic absorption peak of DMPO-•OH, and the presence of DMPO- (1:1:1:1:1:1 six signals) is also captured, indicating that both •OH and exist in the system. In addition, TEMP was used to verify the existence of ¹O₂ in the system, as shown in Figure 5(c). When only PDS is added without steel slag, there was no obvious absorption peak. However, after SA-SSP was added (5 min), a triple signal of 1:1:1 peak-to-height ratio appeared, which was a typical absorption peak of TEMP-¹O₂, indicating the presence of ¹O₂. Therefore, the free radical quenching experiment and EPR fully indicate that •OH and ¹O₂ are mainly presented in the SA-SSP/PDS system.

The activation mechanism of PDS was investigated by using XPS spectra of SA-SSPs before and after the treatment. The full XPS spectrum shows the presence of Fe(2p), Mn(2p), Ca(2p), Si(2p), O(1s), and C(1s) on the catalyst surface, as shown in Figure 5(d). To investigate the relative changes in Fe content at different valence states, the high-resolution XPS spectra of Fe(2p) before and after the use of SA-SSP were fitted (Figure 5(e)). The binding energy peaks at 710.1 and 723.7 eV and satellite peaks at 715.8 eV are ascribed to Fe(Ⅱ), while the peaks at 711.8 and 725.4 eV are ascribed to Fe(Ⅲ) (Xin et al. 2021). By comparing the ratio of Fe(Ⅱ)/Fe(Ⅲ) before and after use, it was found that the ratio of Fe(Ⅱ)/Fe(Ⅲ) decreased from 1.39 to 0.99 after use, indicating that the content of Fe(Ⅱ) decreased, which may be caused by the consumption of Fe(Ⅱ) to activate PDS and converting into Fe(Ⅲ). In addition, SA-SSP also contains a small amount of transition metal Mn, which also plays a certain role in promoting the activation of PDS. The high-resolution XPS spectra of Mn(2p) before and after the use of SA-SSP are shown in Figure 5(f). The peaks of binding energy at 641.0 and 652.2 eV indicate the presence of Mn(Ⅱ). The peaks at 642.3 and 653.5 eV were attributed to Mn(Ⅲ). The peaks at 644.1 and 655.3 eV are Mn(Ⅳ). The relative content of Mn in different valence states changed before and after use; Mn(II):Mn(III):Mn(IV) changed from 0.41:0.28:0.31 to 0.39:0.34:0.27 after the reaction. The relative content of Mn(Ⅱ) decreased and Mn(Ⅲ) increased, indicating that some Mn(Ⅱ) participated in the activation process of PDS, and then was oxidized to Mn(Ⅲ) by PDS and produced reactive oxygen species at the same time.

Liquid chromatography mass spectrometry (LCMS) is an analytical technology that combines the high separation capacity of liquid chromatography and the high sensitivity of mass spectrometry, which not only improves the efficiency and accuracy of the analysis, but also greatly expands the field of chemical analysis and scientific research, and is an indispensable tool in modern chemical analysis and scientific research. In order to analyze the above degraded organic intermediates, 14 intermediates with mass charge ratios (m/z) of 445, 449, 431, 461, 432, 353, 417, 388, 336, 319, 274, 259, 150, and 114 were obtained by LC-MS detection and analysis, as shown in Supplementary material, Figure S3. In addition, in order to better predict the degradation pathway of CTC by the SA-SSP/PDS system, Fukui functions were introduced to predict the active sites of CTC molecules that were the most likely to react with free radicals. Atoms at different sites of CTC molecules had different Fukui indices (f⁻ and f⁰) with large f⁻ values, as shown in Supplementary material, Table S1. This indicates that the higher the electron cloud density on the atom, the more likely it is to react with free radicals. For example, Cl55 (f⁻ = 0.1076), N44 (f⁻ = 0.1428), and O9 (f⁻ = 0.1031) have higher f– values, so the atoms at this site may be more likely to react with reactive oxygen species. Based on Fukui function predictions combined with LC-MS characterization analysis, two possible CTC degradation pathways were summarized (Supplementary material, Figure S4). Pathway Ⅰ: Due to the high density of chlorine substituted electron cloud at Cl55 of the CTC molecule (f⁻ = 0.1076), reactive oxygen species first began to attack from this place and formed intermediate P1 (m/z = 445) by dechlorination, which was consistent with the predicted results of Fukui function. The removed chlorine substituents are then hydroxylated to form P4 (m/z = 461); In addition, dimethyl ammonia in the P1 molecule also has a higher charge density, and the two methyl groups in the N atom can generate P3 (m/z = 431) and P7 (m/z = 417) intermediates by the way of removing monomethyl and dimethyl, respectively. Finally, the hydroxyl group (–OH) at O53 and the amino group (–NH₂) at N41 of the P7 molecule were removed to form the intermediate P8 (m/z = 338). Pathway Ⅱ: N44 in the CTC molecule also had a high f⁻ value (f⁻ = 0.1428). Therefore, dimethylamine is easy to be attacked by active substances and form an intermediate product P2 (m/z = 449). Then, the –OH on O36 in P2 forms P5 (m/z = 345) by dehydration, and then P9 (m/z = 336) is formed by dechlorination, dehydroxylation, and deamidation. In addition, P2 can form P6 (m/z = 353) through demethylation, hydroxyl and amino pathways, and finally, P10 (m/z = 319) through dechlorination. The above organic intermediates reacted with reactive oxygen species to form the rate-opening products P11 (m/z = 274), P12 (m/z = 165), P13 (m/z = 150) and P14 (m/z = 114) with smaller molecular weights. Finally, the organic intermediates were gradually mineralized into H₂O, CO₂, and inorganic salts under the continuous attack of reactive oxygen species.

In this study, salicylic acid was used to modify SSP to obtain a heterogeneous catalyst SA-SSP. The results showed that salicylic acid can selectively remove the calcium silicate mineral phase from SSP, exposing more Fe and Mn metal catalytic active sites. The optimal dosages of SA-SSP and PDS were 1.0 g/L and 2 mM, respectively. When the initial pH was 3, the removal rate of CTC could reach 93.42% within 90 min. Increasing temperature is beneficial to the removal of CTC, and the degradation process conforms to pseudo first-order kinetics. The presence of and in aqueous solution had adverse effects on the removal of CTC, and HA also slightly inhibited the removal of CTC. The prepared catalyst can be effectively separated by magnetic separation and the removal rate of CTC remains at about 80% after four cycles. The results of XPS showed that the conversion of Fe(Ⅱ) to Fe(Ⅲ) was the main reason for the free radical generation of PDS. Free radical scavenging experiments and EPR fully demonstrated that •OH and ¹O₂ were the main reactive oxygen species in the SA-SSP/PDS system. The TOC removal rate was only 46.05%, indicating that there were still a large number of intermediates after degradation. With the progress of the reaction, a part of the intermediates was mineralized into CO₂, H₂O, and inorganic salts.

This work was supported by Major Science and Technology Projects in Yunnan Province (202302AG050002-4); National Natural Science Foundation of China (22466024, 22366023); Yunnan Ten Thousand Talents Plan Young and Elite Talents Project (YNWR-QNBJ-2018-388, YNWR-QNBJ-2020-063).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.