High-purity monoclinic pyrrhotite derived-from natural pyrite with excellent removal performance for Cr (VI) and its mechanism

Pyrrhotite, especially the monoclinic type, is a promising material for removing Cr (VI) from wastewater and groundwater due to its high reactivity. However, the purity of the preparation monoclinic pyrrhotite from heated natural pyrite is not high enough and the role of possible sulfur vacancies in pyrrhotite’s crystal structure has been largely ignored in the removal mechanism of Cr (VI). In this work, we characterized the phase composition changes of annealed pyrite in inert gas and prepared high-purity (~96%) monoclinic pyrrhotite at the optimal condition. We found that it could remove 18.6 mg/g of Cr (VI) by redox reaction, which is the best value reported of natural pyrite derived materials so far. As the reactive media material of simulated permeable reactive barrier, the service life of the high-purity monoclinic pyrrhotite column is 297 PV, which is much longer than that of the pyrite column (50 PV). A new founding is that S 2-and S vacancy play the essential role during the redox reaction of pyrrhotite and Cr (VI) . Monoclinic pyrrhotite had more S vacancy than hexagonal pyrrhotite and pyrite, which explained its superior Cr (VI) removal performance.


Introduction
Water pollution by chromium (Cr) is a common environmental issue.Cr can be found in water as either trivalent (Cr (III)) or hexavalent (Cr(VI)) forms.Cr(VI) is very harmful to living beings, as it can cause cancer, skin ulcers, kidney damage and other problems.In contrast, Cr (III) is less toxic and more soluble.
Therefore, removing Cr(VI) from wastewater/groundwater or changing it to Cr (III) is a crucial and urgent task.
Iron sul des have been explored as functional materials of permeable reactive barriers for removal of Cr(VI) in contaminated groundwater.Pyrite and pyrrhotite, as two common iron sul des, have important roles in geochemistry and technological processes (Cody et al. 2004;Huang et al. 2015;Ma et al. 2016).Due to their strong reducing capacity and a nity for heavy metals, iron sul des are often used as cheap and effective materials for reducing radionuclides and adsorbing heavy metal elements in wastewater or contaminated groundwater treatment (Duan et al. 2016; Erdem and Ozverdi 2006;Li et al. 2014;Weisener and Gerson 2000).Pyrite (FeS 2 ) is a cubic crystal with a NaCl-type structure, belonging to the space group Pa3.It is non-magnetic and has a relatively low reactivity in chemical reactions (Rickard and Luther 2007;Zhu et al. 2019).Pyrrhotite (Fe 1 − x S, 0 < x < 0.125, from FeS to Fe 7 S 8 ), is typically a defective compound with a disordered NiAs structure due to its non-stoichiometric composition, and is known to have a number of "polytypic" or ordered supercell forms that fall into two basic subgroups: Monoclinic pyrrhotite (Fe 7 S 8 ) and hexagonal pyrrhotite (Fe 9 S 10 ).They are derived from metamorphic and sedimentary pyrrhotite, respectively (Dorogina et al. 2015;Janzen et al. 2000).Pyrrhotite, especially the monoclinic type, is more reactive than pyrite because of its iron de ciency in the crystal structure ( The application of natural iron sul de minerals is limited by their crystal structure and size.In contrast, synthetic iron sul de nanomaterials have enhanced reactivity due to the nano effect (Jeong et al. 2010).
However, they also have drawbacks such as high preparation cost and poor stability, which hinder their practical application (Chen et al. 2014;Erdem et al. 2001).Alternatively, pyrite can be transformed into a mixture of monoclinic and hexagonal pyrrhotite by annealing in inert gas under suitable conditions (Zhou et al. 2023), and the type of pyrrhotite formed depends mainly on the annealing temperature (Bhargava et al. 2009).Previous studies have shown that the pyrrhotite obtained from annealed pyrite has a good removal e ciency for heavy metals such as mercury (Liao et (Chen et al. 2015), and so on (Yang et al. 2017), in wastewater treatment, and its reactivity is higher than that of natural pyrrhotite.Moreover, monoclinic pyrrhotite tends to have better removal performance than hexagonal pyrrhotite (Lu et al. 2006;et al. 2016).Therefore, it is desirable to obtain a high-purity monoclinic pyrrhotite product from annealed pyrite for enhanced heavy metal removal.
However, the monoclinic pyrrhotite content in the products was generally low (up to 50%) (Zhou et  In this work, we rst investigated the phase transformation of natural pyrite under different annealing conditions in inert atmosphere by using XRD Rietveld re nement, and obtained a high-purity (~ 96%) monoclinic pyrrhotite sample.Its Cr (VI) removal capacity reached to 18.6 mg/g, which is the highest value reported of natural pyrite derived materials so far.We further explained the reasons for the excellent Cr (VI) removal performance of monoclinic pyrrhotite, which were mainly related to more sulfur vacancies in monoclinic pyrrhotite.
2 Material and methods

Materials and reagents
The natural pyrite was collected from Daye, Hubei Province, in southern China.The crude ore was crushed and sieved to obtain a pyrite powder with a particle size of less than 200 mesh.The pyrite powder was then soaked in 5% HCl to remove surface oxides and impurities, and then annealed in a mu e furnace at different temperatures and durations to study the phase transition and the content of each phase in the process.The annealing conditions were as follows: 500, 600, 700, 800°C for 180 min with a heating rate of 10°C /min; and 600℃ for 30, 60, 120, 140, 160, 180 min with the same heating rate.The annealing was performed in an inert atmosphere to prevent oxidation.A stock solution of Cr (VI) was prepared by dissolving 2.8287 g of K 2 Cr 2 O 7 in 1 L of deionized (DI) water.

Characterizations
The phase compositions and crystal structures of the samples were determined by X-ray diffraction (XRD) using an X-ray powder diffractometer (D8 Advance, Bruker, Germany) with Cu Kα1 radiation (λ = 1.5406Å) at 40 kV and 100 mA.The XRD patterns were recorded from 5° to 130° with a scanning speed of 1.2°/min and a step size of 0.02°.The Rietveld re nement method was used to quantify the phase contents of the samples.Scanning electron microscopy (SEM, ZEISS, sigma 300, Germany) was used to observe the morphologies of the samples.X-Ray uorescence spectrometry (XRF, Thermo electron corporation, ARLADVANT X, USA) was used to analyze the chemical composition of the samples.The Brunauer-Emmet-Teller (BET) speci c surface area was measured from N2 adsorption-desorption isotherms using an automated gas sorption instrument (Micro Active for ASAP 2460, USA).The electron paramagnetic resonance (EPR) spectra were obtained on a Bruker EMX plus model spectrometer operating at the X-band frequency (9.4 GHz) at room temperature to investigate the oxygen vacancies in the samples.The valence states of different elements in the samples were examined by X-ray photoelectron spectroscopy (XPS, Thermo Scienti c, ESCALab250, USA).The pH of the solutions was monitored using a pH meter (pHS-3C, China).

Removal of Cr (VI) at batch experiment
The batch experiments of Cr (VI) removal by pyrite and its annealed products were carried out by varying the initial concentration and contact time while keeping the adsorbent dosage constant at 5 g/L.
Standard solutions with different concentrations of 5, 10, 20, 40, 80, 120, 160 and 200 mg/L were prepared by diluting the stock solution, and a volume of 20 mL Cr (VI) solution was used in each batch experiment.The pH of the solution was adjusted to neutral to simulate the tailing leachate of a polymetallic mine in China.The mixtures were shaken on a reciprocal shaker at 150 rpm for a predetermined time ranging from 0.1 to 12 hours.After shaking, the mixtures were centrifuged at 8000 rpm for another 10 min.The residual Cr (VI) concentration in the supernatant was determined by using the 1,5-diphenylcarbazide method at a wavelength of 540 nm(Bartlett and James 1979).

Removal of Cr (VI) at Dynamic column experimental
Different iron mineral materials (20 ~ 50 mesh) were packed into glass columns to simulate permeable reactive barrier (Figure S1).The column diameter was 3.6 cm, the total height was 15 cm, and the experimental column was divided into three different regions from bottom to top: water distribution zone, reaction zone, and water outlet zone, each with a height of 5 cm (Figure S1a).When lling the glass column, a rubber stopper with a glass tube was used at the lower end of the glass column, and a layer of lter cloth was laid on the upper end of the stopper to prevent the particles of the lling medium material from entering the tube and causing blockage.Then, quartz sand with a particle size of about 3 mm was lled, with a lling height of about 5 cm.The quartz sand was evenly lled, and the lling column was compacted continuously during the lling process.After that, a nylon mesh was placed on the quartz sand to prevent the medium material from leaking into the quartz sand, and 5 cm high of medium material was evenly and compacted lled to ensure that the lled medium material was densely stacked in the column.After the medium material was lled, another nylon mesh was placed to separate the medium material, and then quartz sand was lled.Then another layer of lter cloth was placed to prevent quartz sand particles from owing into the upper tube and causing blockage, and also to lter the out owing liquid to prevent particles and powder from making the liquid sample turbid.A 10-head peristaltic pump was used to inject different experimental columns from below, and the e uent was collected at xed intervals and measured for pollutant concentration (Figure S1b).When breakthrough occurred in the experimental column, the solution was switched to deionized water for desorption.
3 Results and discussion

Characterization and phase composition changes of the products of heat-treated pyrite in inert atmosphere
To investigate the phase transformation of heat-treated pyrite in argon atmosphere, we annealed the pyrite powder at different temperatures and durations and analyzed the phase and content changes of the products by using in-situ and ex-situ XRD techniques.Figure 1 shows the in-situ XRD contour pro le of the products of heat-treated pyrite at different annealing temperatures.The diffraction peaks at 28.6°, 33.2°, 37.2°, 40.8°, 47.4° and 56.4° correspond to the (111), ( 200), ( 210), ( 211), ( 220) and (311) planes of pyrite, respectively.When the annealing temperature reaches about 500℃, new diffraction peaks at 30.0°, 34.0°, 44.0° and 53.3° emerge, corresponding to the ( 22), (004), (322) and (040) planes of pyrrhotite, respectively.When the annealing temperature reaches about 600℃, the diffraction peaks of pyrite completely disappear.The main difference between monoclinic pyrrhotite and hexagonal pyrrhotite in the XRD pattern is that monoclinic pyrrhotite has multiple diffraction peaks while hexagonal pyrrhotite has fewer diffraction peaks in the range around 44° of 2θ (Horng 2018).However, it is di cult to distinguish them clearly in the in-situ XRD patterns because of the relatively fast scanning rate during the measurement.Therefore, we performed ex-situ XRD measurements with a much lower scanning rate and quantitatively analyzed the results by using the Rietveld re nement method.These re ned results are shown in Figure S2, Table S1 (Supporting information) and Table 1.The sample obtained at 600℃ for 1 hour (Pyr-600-60) contains 56.87% pyrite and 43.13% monoclinic pyrrhotite, while the sample obtained at 700℃ for 1 hour (Pyr-700-60) is pure hexagonal pyrrhotite.When the annealing temperature increased to 800℃ for 1 hour (Pyr-800-60), the nal product is pure troilite.Moreover, the content of pyrite gradually decreased and the content of monoclinic pyrrhotite gradually increased at 600℃ with increasing annealing duration (Fig. 1e and Table 1).When the duration was 160 min (Pyr-600-160), the content of monoclinic pyrrhotite reached the highest (~ 96%).When the duration reached 180 min (Pyr-600-180), all the pyrite in the sample disappeared, and hexagonal pyrrhotite began to coexist with monoclinic pyrrhotite.Thus, the optimal condition for obtaining high-purity monoclinic pyrrhotite is annealing at − 1 600℃ for 160 min, and the optimal condition for obtaining pure hexagonal pyrrhotite is annealing at 700℃ for 60 min under argon atmosphere.To further understand the morphology and composition of pyrite and high-purity monoclinic pyrrhotite, we performed SEM, EDS and ICP analyses on the raw pyrite and Pyr-600-160 samples.Figures S3 shows digital photographs of pyrite, high-purity monoclinic pyrrhotite and hexagonal pyrrhotite respectively.It can be seen that the raw pyrite ore is golden in colour with a strong metallic luster.The high-purity monoclinic pyrrhotite is dark bronze in colour and the hexagonal pyrrhotite is dark brown in colour, both of which have no metallic luster.The SEM images show that there is no signi cant change in the particle size and surface morphology of pyrite before and after annealing (Fig. 2a-2b).The EDS results indicate that the main elements in pyrite are iron and sulfur, and there are almost no other impurity elements.Moreover, the S element mapping shows that the sulfur content in pyrite is higher than that in pyrrhotite (not shown).The ICP results of pyrite, monoclinic pyrrhotite and hexagonal pyrrhotite are shown in Table S1 (Supporting information).The sulfur contents in high-purity monoclinic or hexagonal pyrrhotite are signi cantly lower than that in pyrite.This is consistent with the observation that a large amount of elemental sulfur was deposited on the inner wall of the tube furnace during the annealing process, con rming the loss of sulfur in the pyrrhotite samples.
To investigate the effect of sulfur precipitation on the speci c surface area of the samples during annealing, we measured the speci c surface area of pyrite, monoclinic pyrrhotite and hexagonal pyrrhotite samples by using BET method.Figure 2 shows the nitrogen adsorption and desorption curves of the three samples.According to the IUPAC classi cation, the adsorption/desorption isotherms of the three samples all belong to type II, indicating that they have mesoporous structures formed by capillary condensation.The speci c surface area and pore parameters of the three samples are shown in Table S2 (Supporting information).The speci c surface area slightly increases from 0.723 m 2 /g for pyrite to 1.338 m 2 /g for monoclinic pyrrhotite or 1.042 m 2 /g for hexagonal pyrrhotite after annealing.However, the pore volume and pore diameter show little difference among the three samples.

Cr (VI) removal performance of the products of heattreated pyrite at batch and column experiment
Figure 3a shows the Cr (VI) removal capacities of the products of pyrite annealed at different temperatures for 1 hour.The Cr (VI) removal capacity is obviously improved after annealing of pyrite.
When the annealing temperature is 600℃, the Cr (VI) removal capacity is the highest, reaching 2.97 mg/g.Then, we investigated the Cr (VI) removal capacities of the products of pyrite annealed at 600℃ for different durations, and the results are shown in Fig. 3b.When the duration is 160 min, the Cr (VI) removal capacity is the highest, about 4.45 mg/g.If the duration is increased further, the Cr (VI) removal capacity decreases.Therefore, the sample prepared at 600℃ for 160 min has the best Cr (VI) removal performance.It is obvious that the Cr (VI) removal capacity of the products of heat-treated pyrite is highly correlated with the content of monoclinic pyrrhotite.
Based on the above preliminary experiment, we selected pyrite, high-purity monoclinic pyrrhotite (pyrite-600-160) and pure hexagonal pyrrhotite (pyrite-700-60) as the main research materials.We studied the effects of initial concentration of Cr (VI) on their removal performance, and the results are shown in Fig. 3c.The maximum Cr (VI) removal capacities for pyrite, high-purity monoclinic pyrrhotite and hexagonal pyrrhotite are 1.2 mg/g, 18.6 mg/g and 2.5 mg/g, respectively.To evaluate the removal performance of these materials, we compared them with some related literature data.In general, natural pyrite has poor Cr (VI) removal capacity, and the maximum removal amount is about 0. ).When the annealing temperature increased to 700℃, the removal rate decreased from 99.7-13.8%.Therefore, the high-purity monoclinic pyrrhotite prepared in our work has the best Cr (VI) removal capacity reported of natural pyrite derived materials so far.In addition, it should be noted that the pH value of the reaction solutions were adjusted close to 7 in order to simulate the groundwater environment in our work.The removal capacity of Cr (VI) or many other heavy metals of metal sul des can be greatly improved in the acidic reaction solutions, which has been discussed in previous work (Lu et al. 2006;Yang et al. 2017).
To evaluate its application potential in practical permeable reactive barriers, we conducted a column experiment to simulate groundwater environment with an initial concentration of 5 mg/L of Cr (VI).
Figure 3d shows the corresponding experimental result of column experiment.When the e uent concentration ratio is de ned as 0.5, the corresponding pore volume (PV) is taken as the service life of the column.The service life of the pyrite column is only 50 PV, indicating that pyrite has a relatively poor Cr (VI) removal performance.In contrast, the service life can reach 297 PV for high-purity monoclinic pyrrhotite under the same condition, suggesting its excellent durability.

Crystal structure and defects
We studied the crystal defects of pyrrhotite and the mechanism of Cr (VI) removal by using composition analysis, Rietveld re nement, EPR and XPS methods.First, we calculated the chemical formula from the elemental composition of the minerals.According to the composition analysis results in Table S2 (Supporting information), the chemical formulas were Fe 0.999 S 2 , Fe 0.876 S and Fe 0.900 S for pyrite, monoclinic pyrrhotite and hexagonal pyrrhotite respectively, assuming that there was no S vacancy.Pyrite had almost no crystal defects.In contrast, some Fe 2+ was oxidized to Fe 3+ in pyrrhotite, resulting in iron vacancies.After considering the charge balance, the chemical formula of monoclinic pyrrhotite can be written as Fe 2 + 0.628 Fe 3+ 0.248 S 2− , i.e., Fe 7.008 S 8 , which is consistent with the structural formula Fe 7 S 8 proposed by previous studies.The chemical formula of hexagonal pyrrhotite was Fe 2 + 0.700 Fe 3+ 0.200 S 2− , i.e., Fe 9 S 10 , which is consistent with the presumed structural formula Fe 9 S 10 .It can be seen that more Fe 2+ is oxidized to Fe 3+ in monoclinic pyrrhotite, so it has more iron vacancies than hexagonal pyrrhotite, con rming the previous speculation (Shi et al. 2016).It is generally believed that monoclinic pyrrhotite has better Cr (VI) removal performance than hexagonal pyrrhotite due to more Fe vacancies in monoclinic pyrrhotite (Lu et al. 2006;et al. 2016).However, there are no reports on the sulfur vacancy of monoclinic pyrrhotite and hexagonal pyrrhotite.Interestingly, we found that S vacancy also exists in monoclinic and hexagonal pyrrhotite besides Fe vacancy in our work through XRD Rietveld re nement and EPR measurements.Figure 4 and Table S3 show their re ned crystal structure diagrams and lattice parameters.The occupancies of Fe and S atoms in these three samples are shown in Table S4 (Supporting information).The occupancy of Fe in monoclinic and hexagonal pyrrhotite is 0.870 and 0.887 respectively, which is consistent with the previous structural formulas of these two samples.S vacancy is also found in monoclinic and hexagonal pyrrhotite besides Fe vacancy, and monoclinic pyrrhotite contains more S vacancy than the other two samples.To further con rm the S vacancy in these samples, we performed EPR measurements on pyrite samples annealed at different temperatures.As shown in Fig. 4d, the signal at g = 2.004 is attributed to sulfur vacancy (Wang et al. 2020;Yang et al. 2022;Zhang et al. 2018).The content of sulfur vacancies in these samples increases gradually with increasing annealing temperature of pyrite.When the annealing temperature is 600℃ (the monoclinic pyrrhotite content is the highest), the signal of S vacancy is the strongest.Then, when the temperature increases to 700℃ or 800℃, the content of S vacancy decreases slightly, but it is still higher than that of natural pyrite.These results con rm that monoclinic pyrrhotite has more S vacancy than hexagonal pyrrhotite and pyrite.Based on the above results, we modi ed the chemical formulas of the three samples as Fe 2 + 0.999 S −1 1.994 , Fe 2 + 0.698 Fe 3+ 0.172 S 2− 0.956 and Fe 2 + 0.707 Fe 3+ 0.180 S 2− 0.977 for pyrite, monoclinic and hexagonal pyrrhotite respectively, according to the site occupancies of Fe, S and their charge balance.

Valence states of Fe, S and Cr of the samples before and after Cr removal
To reveal the valence states of iron and sulfur elements in pyrite and monoclinic pyrrhotite and the change of valence states after the adsorption of chromium by monoclinic pyrrhotite, we performed XPS analysis on the corresponding samples.The results are shown in Fig. 5. Figure 5a and 5b are Fe 2p and S 2p spectra of pyrite, respectively.The Fe 2p spectrum of pyrite shows two main peaks at 707. 5

Conclusions
In this work, we investigated the thermal evolution of natural pyrite from natural pyrite by annealing in argon atmosphere at different temperatures and durations, and nally prepared high purity monoclinic pyrrhotite (~ 96 wt%).The high-purity monoclinic pyrrhotite has the highest Cr (VI) removal capacity of 18.6 mg/g, which is the best value reported of natural pyrite derived materials so far.The service life of the high-purity monoclinic pyrrhotite column is 297 PV, which is much longer than that of the pyrite column (50 PV) when they used as the reactive media material of the simulated permeable reactive barrier.The Cr (VI) removal mechanism by pyrrhotite is a process of dissolution and reduction.S 2− plays the essential role in the redox reaction of pyrrhotite and Cr (VI) system.In addition, S vacancy also contributes to the Cr (VI) removal by attracting CrO 4 2− in the solution and exposing more active reaction sites.Monoclinic pyrrhotite has more S vacancy than hexagonal pyrrhotite and pyrite, which explains its superior Cr (VI) removal performance.Therefore, our work demonstrated that high purity monoclinic pyrrhotite synthesized from heat-treated natural pyrite has great application prospects as an e cient and low cost active material in permeable reactive barriers for removing of Cr pollutant.The re nement structures of (a) pyrite, (b) monoclinic pyrrhotite and (c) hexagonal pyrrhotite.The yellow sphere is the iron atom and the silver sphere is the sulfur atom.

Table 1
The content of each phase in samples prepared under different treatment conditions (Luo et al. 20079)ural pyrite indicates partial oxidation on the surface, which is consistent with literature reports (Yong et al. 2016;Zheng et al. 2019).Figure5cand 5d are Fe 2p and S 2p spectra of pyrrhotite, respectively.The Fe 2p 3/2 spectrum of pyrrhotite shows four peaks at 707.5 eV, 711.4 eV, 721.0 eV and 725.4 eV, which belong to Fe 2+ bonded with S 2− and Fe 3+ bonded with -OH, respectively.It can be clearly seen that after pyrite annealing, the content of Fe 3+ in the formed pyrrhotite increases signi cantly, replacing part of Fe 2+ , thus creating iron vacancies in the monoclinic pyrrhotite.The S 2p spectrum of pyrrhotite shows two main peaks at 161.2 eV and 162.6 eV, which are attributed to VI) in solution to Cr (III), it was adsorbed on the surface of pyrrhotite as Cr 2 S 3 precipitate(Luo et al. 2007).However, based on the analysis of XPS results, we found that S 2− plays the essential role rather than Fe 2+ in the redox reaction of pyrrhotite and Cr (VI) system in our work.In addition, the Cr (VI) removal capacity highly relates to the content of S and Fe vacancies in the annealed samples (monoclinic pyrrhotite > hexagonal pyrrhotite > pyrite).Vacancy engineering is an effective strategy to tune interfacial interactions to promote the dynamic properties of charge carriers, which are widely investigated in the elds of photocatalyst(Zhang etal.2023) and energy storage (Zheng et al. 2021), and so on.The S vacancies can introduce positive charge centers in the local structure (Zhang et al. 2023; Zheng et al. 2021), which can attract CrO 4 2− in the solution and facilitate the redox reaction.From this perspective, S vacancy may play a more important role than Fe vacancy during the redox reaction process.Therefore, based on the above discussions, we propose a reaction mechanism as shown in Fig. 6.After annealing, the S 2 2− of pyrite is partly oxidized to elemental sulfur and partly reduced to S 2− , and some S vacancies are generated in the structure.During the redox reaction, Cr (VI) is reduced to Cr (III) and combined with dissolved S 2− to form Cr 2 S 3 precipitate.At the same time, S 2− on the surface of pyrrhotite is oxidized to SO 4 2− .In addition, monoclinic pyrrhotite has the most S vacancies, which expose more active reaction sites and promote the redox reaction.As a result, monoclinic pyrrhotite has the best Cr (VI) removal performance.
(Pettine et al. 1998of microsolubility, reduction and precipitation.Therefore, iron sul de can react with variable valence metal ions via redox reactions(Yanming et al. 2010).In the iron sul de-Cr (VI) system, Fe 2+ in the system has a strong reducing effect, which may reduce Cr (VI) in solution.At the same time, S 2− may have a stronger reducing effect than Fe 2+(Pettine et al. 1998) and also can reduce Cr (VI) to Cr (III).Previous studies suggested that these two types of reactions occurred simultaneously.After reducing Cr (