New insight into mechanism of Cr(VI) migration and transformation in typical soils of chromite ore processing residue (COPR) contaminated sites

Chromite ore processing residue (COPR) storage sites are widely distributed all over the world, causing serious soil and groundwater pollution. However, the differences in soil constituents and properties in different regions are significant, and the dynamic migration and transformation of Cr(VI) in different types of soil under alkaline condition of the COPR site is still unclear. In this study, typical black soil, red soil and loess in different regions of China were chosen to investigate the adsorption kinetics and thermodynamics of Cr(VI) under the original pH conditions of the soil, and then the alkaline Cr(VI) solution was introduced into the soil column to simulate the dynamic migration and transformation process of Cr(VI) at COPR sites. According to the results, the Cr(VI) breakthrough curve predicted by the solid–liquid distribution coefficient Kd based on the static isotherm adsorption experiments significantly underestimated and overestimated the retention effect of black soil and red soil on Cr(VI) dynamic migration, respectively. For the black soil, the retention of Cr(VI) was dominated by Cr(VI) reduction, which was a slow reaction compared with Cr(VI) adsorption. Therefore, the reduction kinetics process during the column experiment cannot be neglected. With respect to the red soil, the outlet Cr(VI) concentration turned to be higher than the inlet concentration with the soil alkalization, which indicated that the adsorbed Cr(VI) desorbed again, and this was the main reason for the overestimation of Cr(VI) retention effect by the red soil. This study shows that the environmental risks of Cr in different types of soil are quite different, mainly related to the valence and occurrence form of Cr governed by the soil constituents and properties. In addition, the stable form of Cr in the black soil column after the reaction indicates that the soil organic matter can be used as a potential environmentally friendly remediation material for Cr(VI) contaminated soils at COPR sites.


Introduction
Chromium salt is an important chemical raw material widely used in electroplating, tanning, metallurgy, ceramics, dyes and wood preservation (Dhal et al. 2013). However, it is worth noting that hexavalent chromium (Cr(VI)) has potential cancer risk and is listed as one of the priority control heavy metal pollutants by USEPA Superfund (Saha et al. 2011). In the past few decades, the high lime roasting technology used in the production of chromium salts has produced a large amount of chromite ore processing residue (COPR), and many COPR storage sites have been formed worldwide. About 0.04-2.59% (w:w) of Cr(VI) remains in the COPR, and since a large amount of alkaline substances such as calcium carbonate remain, the COPR tends to exhibit a strong basic character (Chrysochoou et al. 2009;Foeldi et al. 2013). Cr(VI) exists in the form of CrO 4 2− under alkaline conditions and has strong mobility (Zhang et al. 2018c). It is easy to infiltrate into the groundwater with the leaching filtrate of the precipitation, and spread to the downstream.
China is the largest producer of Cr slag (Gao and Xia 2011). Since 1958, 63 chromium salt production plants have been built, covering 23 provinces, municipalities and autonomous regions, involving a variety of soil types, including black soil in the northeast, loess in the Loess Plateau and red clay in the southwest. The accumulated amount of Cr slag over a span of 30 years was more than 6 million tons ). The soil and groundwater around the COPR storage site are generally heavily contaminated, which has become a secondary pollution source that continues to pose a threat to the surrounding ecological environment and human health (Wang et al. 2011).
The migration and transformation process of Cr(VI) in soils is mainly controlled by adsorption/desorption, oxidization/reduction and dissolution/precipitation (Fonseca et al. 2009;Jardine et al. 1999). Cr(VI) will be reduced to Cr(III) when soil organic matter, S(II) and Fe(II) exist (Fendorf 1995). The generated Cr(III) can be complexed by soil organic functionalities or directly form hydroxide precipitation (Zhang et al. 2017(Zhang et al. , 2018a. The mobility of Cr(III) in soil is generally poor, but migration also occurs under conditions of low soil pH or forming complexes with dissolved organic matter (DOM) (Weng et al. 1994). In addition, when MnO 2 minerals are present in the soil and the organic matter content is low, Cr(III) may also be oxidized to Cr(VI) (Landrot et al. 2012a, b). The retention process of Cr(VI) migration in soil involves both specific adsorption and nonspecific adsorption. Non-specific adsorption can be taken as electrostatic adsorption. The specific adsorption mainly refers to the ligand exchange reaction between Cr(VI) and the surface ligand of soil particles. The common ligands include soil humus functionalities and iron oxyhydroxide surface hydroxylic groups (Jiang et al. 2008).
The migration and transformation processes of Cr(VI) in soil is closely related to soil material composition and physicochemical properties (Kožuh et al. 2000;Zhang et al. 2019a). The inorganic minerals in the soil have different degrees of adsorption and reduction of Cr(VI). Iron oxyhydroxides, such as hematite, goethite and ferrihydrite, have a strong adsorption effect of Cr(VI) at an acidic condition, which mainly form inner-sphere complexes Chrysochoou 2012, 2014;Xie et al. 2015). On the contrary, the adsorption of Cr(VI) by aluminum oxides, such as boehmite and gibbsite, is mainly dominated by outer-sphere complexes by electrostatic attraction adsorption and ion exchange adsorption Chrysochoou 2015, 2016). When Fe(II) is contained in the iron oxide, the iron oxide has a reducing ability on Cr(VI), such as magnetite (He and Traina 2005;Jiang et al. 2014). Cr(VI) anions are generally weakly adsorbed on clay minerals because the surface is usually negatively charged (Bradl 2004).
Soil organic matter has a significant retention effect on the migration of Cr(VI). This is because the surface of soil organic matter contains a large number of reactive functional groups, which can complex with Cr(VI) or reduce Cr(VI) into Cr(III). Dissolved organic matter can reduce Cr(VI) at low pH, and the removal efficiency greatly decreases with the increasing pH (Gong et al. 2015). Dissolved organic matter can also form water-soluble complexes with Cr(VI) and Cr(III) to enhance their mobility (Weng et al. 2002). The undissolved organic matter can provide active sites for the adsorption of Cr(VI) and electrons for the reduction of Cr(VI), which plays a major role in the retention of Cr(VI) migration (Barnie et al. 2018;Zhang et al. 2019b;Zhang et al. 2018b). In addition, pH also has an important influence on the migration of Cr(VI). Generally, the lower pH is better for the reduction and adsorption of Cr(VI) (James and Bartlett 1983).
It is worth noting that different types of soil have various components. And the retention effects on Cr(VI) migration are significantly different (Fernandez-Pazos et al. 2013;Lee et al. 1999). However, the retention effects of the typical soils in China on Cr(VI) migration in the COPR sites have not been reported yet. It has been proven that water in contact with COPR had a pH of 11-12, and can contain high concentration of Cr(VI) as the highly mobile and toxic anion chromate (Geelhoed et al. 2003). The migration and transformation of Cr(VI) in soil under hyperalkaline condition may be quite different from the neutral or acidic conditions. In this study, black soil, red soil and loess from different areas of China were chosen to investigate the dynamic transportation and transformation of Cr(VI) in typical soils under extremely basic condition of COPR site, and the variation of Cr speciation in soils were investigated as well. This study will contribute to the development of Cr(VI) ecological risk assessment and remediation schemes design for COPR sites.

Soil samples
To explore the migration and transformation characteristics of Cr(VI) in the typical black soil area, red soil area and loess area in China, the black soil of Antu City, Jilin Province (42°57′54″ N, 128°41′27″ E), the red soil of Qujing City, Yunnan Province (24°59′57″ N, 103°36′21″ E), and the loess of Shangluo City, Shanxi Province (33°29′41″ N, 110°52′15″ E) were collected as the experimental soil samples. And these soil samples are all background soils about 500 m away from the contaminated sites. When collecting soil samples, the areas with little human activity were selected. The surface vegetation was removed and soil within 30 cm below the surface was collected. After removing the crushed stone, plant debris and other impurities in the soil samples, the soil samples were naturally air-dried. Then, the soil samples were ground and sieved to less than 0.25 mm in diameter, and stored in a dry and cool place for later use.
The chemical compositions of the soil samples were determined by X-ray fluorescence (XRF) using a Rigaku ZSX Primus spectrometer. The soil organic matter contents of the soil samples were determined by the muffle furnace method (Klingenfuss et al. 2014). And the soil samples were heated at 550 °C for 4 h. X-ray diffraction (XRD) patterns were obtained in a Rigaku D/max-RA powder diffractionmeter, which was used to determine the mineral compositions and clay mineral compositions of the soil samples. The point of zero charge (PZC) of the soil samples were determined by the salt titration method (Sijin et al. 2006). The pH of the soil samples were measured in water and 0.01 M CaCl 2 at a soil/solution ratio of 1: 2.5 using a pH meter (Sartorius, PB-10) after calibration (Choppala et al. 2013).

Adsorption kinetics and thermodynamic experiment
The adsorption kinetics experiment was carried out in a 50-mL polyethylene centrifuge tube. A Cr(VI) stock solution (1000.0 mg/L) was prepared by adding 2.829 g dried potassium dichromate (K 2 Cr 2 O 7 ) into 1 L of deionized water and then stored at 4 °C. The Cr(VI) solution (10.0 mg/L) was prepared by further diluting the stock solution using deionized water. 1 g of black soil, red soil and loess samples were added, separately. Then 30 mL of 10 mg L −1 Cr(VI) solution was added. 0.02 M NaCl was used as the supporting electrolyte to simulate the ion strength in actual groundwater. The initial pH was not adjusted, which was in accordance with the pH of the deionized water (about 6.5). The tubes were placed in a constant temperature shaker at 25 °C, 175 rpm for horizontal shaking. The samples were taken out at regular intervals, and the maximum reaction time was 24 h. The tubes were centrifuged at 3000 rpm for 10 min, and the supernatant was passed through a 0.45 μm filter. The pH of the filtrate was measured, and the concentration of Cr(VI) in the filtrate was determined by UV-vis spectrophotometer (SHIMADZU UV-1800). Then, the adsorption quantity of Cr(VI) by the soil sample was calculated according to the decreasing Cr(VI) concentration in solution. The experimental group set up three parallel samples with 0.02 M NaCl background electrolyte. The blank sample was designed without Cr(VI), and the control sample was designed without soil samples. The Cr(VI) adsorption thermodynamics experiment was consistent with the adsorption kinetics experiment, except that the samples were taken out after shaking of 24 h and the initial concentrations of Cr(VI) were set to 5-50 mg L −1 . Experiment on the effect of pH on adsorption of Cr(VI) by soil was consistent with the adsorption kinetics experiment, except that the samples were taken out after shaking of 24 h and the initial pH values of Cr(VI) solution were set to 2-12.

Adsorption kinetics and isotherm model analysis
The pseudo first-order and pseudo second-order kinetic equations were utilized to analyze the kinetic experiment data. The pseudo first-order kinetic equation is derived from the equation dc dt = −k 1 c − c e (Moussout et al. 2018). It is shown as follows: where, q t represents the Cr(VI) adsorption quantity on the soil samples. C 0 and C e represent the initial concentration and equilibrium concentration of Cr(VI) in the solution, respectively. k 1 represents the pseudo first-order kinetic constant. V and m represent the volume of the solution and mass of the soil sample, respectively.
The pseudo second-order kinetic equation is derived from Moussout et al. 2018). It is shown as follows: where, k 2 represents the pseudo second-order kinetic constant. The meanings of the other parameters are consistent with Eq. (1).
The Langmuir and Freundlich isotherm equations were utilized to analyze the thermodynamic experiment data, and the Langmuir isotherm equation is shown as follows: where, q s represents the Cr(VI) adsorption quantity on the soil samples. q m represents the maximum Cr(VI) adsorption quantity on the soil samples. C e represents the equilibrium Cr(VI) concentration in solution. K L represents the Langmuir non-linear solid-liquid partition coefficient.
The Freundlich isotherm equation is shown as follows: where, the K F represents the Freundlich non-linear solid-liquid partition coefficient. The meanings of the other parameters are consistent with Eq. (3). (1)

Column experiment of Cr(VI) dynamic transportation
Column experiment was conducted using three parallel plexiglass columns (3.5 cm in internal diameter; 22 cm height). The columns were filled with 2 cm of glass beads, 2 cm of quartz sand, 15 cm of soil (there were separate soil columns for black soil, red soil and loess), 1 cm of quartz sand and 2 cm of glass beads from bottom to top ( Figure S1). The dry heap method was adopted for packing the soils into the column. That is, for the same soil, whenever the same mass of the soil was loaded, the soil was filled evenly to the same height. The sampling ports were positioned at a distance of 9 cm and 14 cm from the bottom of each column. The samples were collected at regular time intervals. Since the inner wall of the plexiglass column was relatively smooth, it was abraded by a sandpaper with a particle size of 25 μm before filling the medium to avoid preferential flow along the inner wall. In all the columns, the solution was loaded in an upflow mode using an adjustable multiport peristaltic pump set (Longer Precision Pump Co., Ltd., BT 100-1F). The pore volume was quantified by measuring the deionized water used to saturate the column. Furthermore, this value was used to calculate the porosity considering the given geometry of the column (Knorr et al. 2016). The porosity of each column and the weight of soil filled in each column are shown in Table S1. The bulk density was calculated by dividing the mass of soil packed into the column by the volume of soil portion in column.
To illustrate the convection-dispersion process of Cr(VI) in the soil columns, a breakthrough experiment was performed using Cl − . That is, 0.02 M NaCl solution was loaded from the bottom to top. Cl − is generally considered as a conservative tracer, which can be used to indicate the convection-dispersion process of Cr(VI) in the soil column. The program CXTFIT (Toride et al. 1995), a nonlinear least squares algorithm, was applied to the experimental results to determine solute transport parameters for the columns. The actual flow velocity v and the diffusion coefficient D of each soil column were obtained, and the results are shown in Table S5.
To simulate the actual pH conditions of the leaching solution of the COPR site, 100 mg L −1 Cr(VI) solution with pH 11.5 and 0.02 M NaCl background electrolyte were introduced into the soil column. And samples were taken out at two sampling ports and outlet holes at regular intervals. The concentration of Cr(VI) in the liquid phase was measured by UV-vis spectrophotometry, and the change in pH was measured. The data of X-ray photoelectron spectroscopy (XPS) obtained by a Kratos Axis Ultra X-ray photoelectron spectrometer were used to characterize the valence of Cr on the soil surface before and after the experiment.

Speciation analysis of Cr in the soil samples
To find out the changes of Cr content and speciation in the soil column before and after the reaction, the modified BCR sequential extraction method was used to study the unreacted soil and the soil samples at 0-5, 5-10 and 10-15 cm in the post-reaction soil column (Žemberyová et al. 2006). The modified BCR sequential extraction method divides the heavy metal form into water-soluble state, weak acid extractable state, reducible state, oxidizable state and residual state. The specific meaning of each form and the extraction reagents and procedures are shown in Table S2.

Convection-dispersion-reaction equation modeling
Based on the convection-dispersion equation, the CXTFIT program was used to fit the Cl − ion and Cr(VI) breakthrough curves in soil column experiments. The convection-dispersion equation is as follows: where, C represents the concentration of the solutes. D represents the dispersion coefficient. v represents the actual velocity. q s and C s are the flux and solute concentration at the input and output, which represent the boundary conditions. R represents the retardation factor. θ represents the porosity. ρ b represents the bulk density of the soil samples. K d represents the solid-liquid distribution coefficient.

Characterization of the soil samples
The characterizations of the three soil samples are shown in Table 1. It can be found that the content of SiO 2 in the red soil sample is lower compared with the other soil samples. But the red soil contains higher content of Al 2 O 3 and Fe 2 O 3 . Mineral composition analysis verified this point. Clay minerals account for about 55% of the mass in red soil, and kaolinite is the main clay mineral. The iron mineral content is about 7% in red soil, mainly composed of hematite. The major clay mineral in black soil and loess is andreattite. The CaO content in loess is high and the mineral composition analysis demonstrated that there was about 15% of calcite in loess. Additionally, organic matter content reaches 13.8% in black soil, while the organic matter content in red soil and loess is lower, only accounting for 0.47% and 0.63%, respectively.
The black soil and red soil samples belong to acidic soils with pH of 4.77 and 5.06, respectively, while the loess is alkaline soil with pH of 8.07. The acidity of the black soil is related to its high organic matter content (Dai et al. 2009). Organic matter contains a large amount of acidic functional groups (Zhang et al. 2018a) and possibly release a large number of protons. Red soil is the main agrotype in Southwest China, it is rich in iron, aluminum oxide with strong acidity and high consistency (Ma et al. 2013). The alkalinity exhibited by the loess is mainly related to its higher calcite content. According to the determined results of PZC as shown in Figure S2, the PZC of black soil and red soil were lower with 3.67 and 4.29, respectively, and the PZC of loess was higher with 8.19. This is because the acidic functional groups rich in organic matter in black soil can provide a large amount of variable charge (Oades et al. 1989), and the clay minerals rich in red soil have a large number of permanent negative charges through isomorphous substitution (Schoonheydt and Umemura 2017). Additionally, the bulk density of black soil was 0.91 g cm −3 , which was quite lower compared with the red soil and loess (1.14 and 1.29 g cm −3 , respectively), and this is mainly related to the higher content of organic matter (Oades et al. 1989).

Kinetic and thermodynamic experiments
To illustrate the adsorption characteristics of Cr(VI) in different soil samples, the adsorption kinetic experiments of Cr(VI) were carried out and the results are shown in Fig. 1. The experimental results show that the adsorption of Cr(VI) by red soil is stronger, while the adsorption of black soil is weaker, and the adsorption of Cr(VI) by loess is hardly observed. During Cr(VI) adsorption by the soil samples, the pH of the solutions kept almost unchanged ( Figure S3).
The adsorption rate of Cr(VI) by red soil was faster, and it basically reached equilibrium in about 4 h, and the maximum adsorption capacity was about 241.36 mg/kg. Since pH of the solution was above 5 during the reaction, which was higher than PZC of the red soil (PZC = 4.29), the surface of the red soil was negatively charged. And there was electrostatic repulsion between the surface of the red soil and the Cr(VI) anion. Therefore, the electrostatic attraction was not likely to be the driving force of Cr(VI) adsorption on the surface of the red soil. In addition, considering that the adsorption of Cr(VI) by kaolinite is weak (Jin et al. 2014), the adsorption of Cr(VI) by hematite by ligand exchange may be the main driving force for its adsorption on red soil (Adegoke and Adekola 2012; Singh et al. 1993).
It is worth noting that the adsorption of Cr(VI) on black soil did not reach equilibrium at the end of the experiment, which is quite different from that of the red soil. The adsorption process can be divided into two stages, fast and slow. This is because the black soil followed the "adsorption-reduction" mechanism for the retention of Cr(VI) as reported by our previous research (Zhang et al. 2017), that is, Cr(VI) is quickly adsorbed to the black soil surface, and the adsorbed Cr(VI) is reduced to Cr(III) by the reducing functional group in the soil organic matter (such as phenolic groups and hydroxylic groups), wherein the reduction process is a rate limiting step of the entire reaction (Zhang et al. 2019a). That is why the adsorption amount of Cr(VI) increases rapidly in the initial stage of the reaction and then rises slowly, which also indicates that the reduction reaction of Cr(VI) by black soil is still going on.
The results of the kinetic experiments were fitted using pseudo first-order and pseudo second-order reaction kinetic models, and the fitting results are shown in Fig. 1a and Table S3. Since the obtained R 2 values were consistent and closer to unity for the pseudo-second order kinetic equation than for the pseudo-first order kinetic equation, the adsorption kinetics can be well described by the pseudo-second order kinetic equation. This indicates that the retention of Cr(VI) on three soil samples is related to both the Cr(VI) concentration and the number of adsorption reduction sites on the soil surface.
To reveal the thermodynamic characteristics of Cr(VI) adsorption of three soil samples, isothermal adsorption experiments were carried out. The results are shown in Fig. 1b. The adsorption of Cr(VI) by red soil is larger, and the adsorption of Cr(VI) by black soil is lower. The adsorption of Cr(VI) by loess is negligible, which is consistent with the results of the kinetic experiments. With the increase of Cr(VI) concentration, the growth trend of Cr(VI) adsorption quantities by black soil and red soil show nonlinear characteristics. The Freundlich isotherm adsorption models had better fitting performances on the experimental results (fitting results are shown in Table S4). According to the Freundlich model, the maximum adsorption capacity of black soil and red soil can reach 247.11 mg kg −1 and 869.45 mg kg −1 , respectively.
The effect of pH on the adsorption of Cr(VI) on the three soil samples is shown in Figure S4. As indicated, the adsorption of Cr(VI) by the red soil and black soil decreased with pH increasing. The adsorption amount of Cr(VI) by the loess is low, and it remains almost unchanged with pH increasing. The adsorption amount of Cr(VI) by red soil decreased rapidly between pH 5-7, while the adsorption amount of Cr(VI) by black soil decreased almost linearly with the increase of pH. In view of the fact that the PZC of hematite is about 7 (Estes et al. 2013), this further confirms that the adsorption of Cr(VI) by the red soil is mainly controlled by the ligand exchange between Cr(VI) and the surface hydroxyl groups of hematite. So the adsorption of Cr(VI) by the red soil is affected by the variable charge of the surface of hematite significantly. The adsorption of Cr(VI) by the black soil is mainly controlled by the reduction process, and the reduction is mainly restricted by the concentration of hydrogen ions (Huang et al. 2012), therefore, the adsorption amount of Cr(VI) by black soil gradually decreases with the increase of pH value.

Cr(VI) dynamic transport in the soil columns
In the loess column, the Cr(VI) breakthrough process was very fast, which was only slightly delayed compared with the Cl − . This is consistent with the result of the weak adsorption of Cr(VI) on the loess observed in the kinetic and thermodynamic batch experiments. It is interesting to find that the black soil had a stronger retardation effect on Cr(VI) migration than that of the red soil in the column experiment. This is contrary to the results obtained in the batch experiments, wherein the adsorption capacity of red soil for Cr(VI) is much stronger than that of the black soil. The solid red line in Fig. 2 represents the Cr(VI) breakthrough curve, which was simulated by the solid-liquid distribution coefficient K d . By comparing the actual breakthrough curve of Cr(VI) with the simulated breakthrough curve of Cr(VI), it can be found that if the K d is used to evaluate the retardation effect on Cr(VI) by soils in COPR contaminated sites, the effect of black soil will be significantly underestimated, and the adsorption capacity of red soil for Cr(VI) will be significantly overestimated.
As mentioned above, the black soil is rich in organic matter and the adsorption kinetic of Cr(VI) by black soil Fig. 1 The a kinetic and b thermodynamic characteristics of Cr(VI) adsorption by black soil, red soil and loess samples at 25 °C. The initial pH of Cr(VI) solution was 6.5, and the background electrolyte was 0.02 M NaCl. The error bar represented Standard Error of Mean. In kinetic experiments, the simulation results were more consistent with the pseudo second-order reaction kinetic model. The solid lines are Freundlich isotherm in (b) followed "adsorption-reduction" mechanism (Zhang et al. 2017). The humus enriched black soil (van Zomeren and Comans 2007) contains a large number of reducing functional groups, such as phenolic and hydroxylic groups Hsu et al. 2009), and thus the Cr(VI) reduction reaction is still in progress (Zhang et al. 2018a). The breakthrough time of Cr(VI) in the black soil column is about 36 days, which indicates that the reduction of Cr(VI) in the soil column lasted for comparable time. In addition, the reduction of Cr(VI) was a hydrogen-consuming reaction (Reduction reactions is shown as follows, Eqs. 7, 8) (Huang et al. 2012). But the black soil contains a large amount of humus, which is rich in acidic functional groups, such as carboxylic groups and phenolic groups. These functional groups have strong pH buffering abilities. Thus, at the end of the experiment, the leachate pH of the black soil column remained at around 8.5 (lower than the pH in the red soil column which exceeded 10), which also allowed Cr(VI) to be reduced in the black soil column (Sillerova et al. 2014;Wittbrodt and Palmer 1995).
By observing Fig. 2b, it can be found that the outlet Cr(VI) concentration of the red soil column appears higher than the inlet concentration, which indicates that Cr(VI) adsorbed on the surface of the red soil desorbed. Because the alkaline Cr(VI) solution was continuously introduced into the red soil column, the acid-base balance of the red soil was broken. When the pH exceeded the PZC of red soil, the excess hydroxide competed for adsorption sites with Cr(VI), which is also the main reason for the overestimation of the retention effect of red soil on Cr(VI). The same phenomena were observed in the sampling ports at 5 cm and 10 cm ( Figure S5). The area of gray and reddish areas in Fig. 2b is related to the adsorption amount and desorption amount of Cr(VI), respectively. Further observation shows that the two areas are almost equal, indicating that almost all adsorbed Cr(VI) was desorbed into the liquid phase. By comparing the adsorption and desorption amounts of Cr(VI) at the three sampling ports, which was calculated by the breakthrough curves ( Figure  S6), it can be found that the adsorbed Cr(VI) and desorbed Cr(VI) are all equivalent in quantity approximately. This further illustrates that the reversible adsorption acts as the retention mechanism of Cr(VI) on the red soil.
In addition, it can be observed in Figure S5(a) that the concentration of Cr(VI) leachate in the three sampling ports did not reach the initial concentration of Cr(VI). This indicates that there is always a liquid phase reduction process of Cr(VI) in the soil column, which may be caused by the redox reaction of dissolved organic matter in the liquid phase with Cr(VI).
To verify the change of Cr content and speciation in the soil columns before and after the reaction, the soil column was divided into upper, middle and lower sections and analyzed by the BCR sequential extraction method, and the results are shown in Fig. 3. It can be seen from the figure that the three soil samples contain a certain amount of Cr, which is the natural background content of Cr in the soil, mainly in the residual state and oxidizable state. And the content is between 30.38 and 71.76 mg kg −1 . The content of Cr in the red soil is the highest. After the reaction, the total content and speciation of Cr in red clay and loess almost kept unchanged. For the red soil, the desorption of Cr(VI) occurs after the pH of the solution changing to alkaline conditions. For the loess, it is mainly due to the weak adsorption of Cr(VI). It is worth noting that the content of Cr in the black soil after the reaction significantly increased, and the Cr content decreased gradually as it approached the outlet. From the change of speciation, Fig. 2 The dynamic transportation of Cr(VI) in the a black soil, b red soil and c loess columns. The initial Cr(VI) concentration was 100 mg L −1 . The initial pH of the inlet solution was 11.5. The background electrolyte was 0.02 M NaCl. The black dash line represents the breakthrough curve of Cl − determined by convection-dispersion equation fitting. The red solid line represents the breakthrough curve of Cr(VI) determined by convection-dispersion equation prediction with the K d obtained from the isotherm model fitting. These curves correspond to the outlet of the column the content of Cr in the oxidizable state is significantly increased, followed by the Cr in the reducible state. These two forms represent Cr combined with soil organic matter and iron oxide, respectively. It can be seen that the speciation analysis results are consistent with the results of Cr(VI) breakthrough experiment.
To further clarify the valence state of Cr adsorbed on black soil, the black soil samples before and after the reaction were characterized by XPS Cr2p. The results are shown in Fig. 4. It can be seen that the signal of Cr before the reaction is very weak, but the signal of Cr after the reaction is significantly enhanced. Through the peak fitting analysis, Cr(III) is dominated, accounting for 80.56% of the total Cr. It can be inferred that the majority of Cr(VI) adsorbed by black soil has been reduced, which is consistent with the speciation analysis results.

The implications for risk assessment and remediation
The pollution process and characteristics of the soil at COPR site are quite different from other Cr-contaminated sites. This is because the alkaline substances in the COPR can cause the alkalization of the soil, which will significantly affect the dynamic migration and transformation of Cr(VI) in soils. In this study, it was found that the red soil has strong adsorption capacity for Cr(VI) under natural soil pH conditions. However, almost all of the adsorbed Cr(VI) was desorbed as the soil pH increased. This indicates that COPR site in the red soil area has strong environmental risks and should be given enough attention. On the contrary, at the end of the experiment, the leaching solution pH of the black soil column reached about 8.5. The adsorption of Cr(VI) became weak, but the adsorption quantity of Cr(VI) in black soil was significant and did not desorb till the end. The speciation analysis and XPS characterization results showed that the Cr in the black soil mainly existed in the form of Cr(III), which indicates that Cr in the contaminated soil at the COPR site in the black soil area has low environmental risk. The environmental risk of Cr in soils is not only determined by the total amount of Cr, but also by the valence and existing form of Cr.
The current remediation projects of the contaminated soil in the COPR site in China mainly use inorganic reducing agent to reduce Cr(VI) to Cr(III), such as ferrous sulfate (FeSO 4 ) and calcium polysulfide (CaS 5 ) (Moon et al. 2008). However, the phenomenon of Cr(VI) re-release may occur in the remediation site (Moon et al. 2007). There are two reasons for this phenomenon. One is that the Cr(III) produced by reduction is unstable, which can be oxidized to Cr(VI) by the electron acceptors such as Mn oxides (Landrot et al. 2012a, b). The other is that the soil particles did not come into contact with the inorganic reducing agent sufficiently, and the adsorbed Cr(VI) by the soil particles is not effectively reduced (Jagupilla et al. 2009). Additionally, the inorganic reducing agent is easily oxidized and deactivated in the air, so the Cr(VI) inside the soil particles is slowly released again. It can be seen that the remediation of Cr(VI) contaminated soils using inorganic remediation agent is not stable. In fact, the use of inorganic remediation agent has the potential to cause secondary pollution. In this study, it was found that the organic matter in black soil had a strong long-term reducing ability to Cr(VI), which could The XPS Cr2p determination of black soil sample before and after reaction(Soil near the inlet of the column). The dash line represents the baseline, and the blue and red line represent the Cr(VI) and Cr(III), respectively. The binding energy of Cr(VI) and Cr(III) were 579.4 (Elio et al. 1988) and 577.3 eV (Jianjun and Qunji 1994), respectively effectively avoid the phenomenon of Cr(VI) re-release after remediation. In addition, soil organic matter can also provide complex adsorption sites for Cr(III) produced by reduction, which is not easy to desorb or oxidize, and greatly reduces the environmental risks of Cr in soil (Kožuh et al. 2000;Masscheleyn et al. 1992).