Enhanced molybdenum(VI) removal using sulfide-modified nanoscale zerovalent iron: kinetics and influencing factors

Although molybdenum (Mo) at low levels is an essential micronutrient for both plants and animals, it has various adverse health effects on aquatic organisms when its concentration exceeds 5 mg/L in water (Moret & Rubio 2003). Subchronic and chronic oral exposures can result in sterility, growth retardation, hypothyroidism, anemia, liver or kidney abnormalities, and death (Namasivayam & Sangeetha 2006). Hexavalent molybdate (⁠⁠) is the most soluble species of Mo (Ma et al. 2015) and is the dominant species in molybdenum-contaminated industrial wastewaters in the manufacture of paints and ceramics, metal alloys, missiles and aircraft, solid lubricants, and pigments for printing inks (Namasivayam & Sureshkumar 2009). Indeed, Mo is widely detected in industrial wastewater, rivers, lakes, and groundwater, and Mo concentration in some waters has seriously exceeded 5 mg/L (Table S1). Therefore, effective technology is needed for removing Mo(VI) from contaminated waters.

Various technologies have been studied for Mo(VI) removal from waters, including chemical precipitation (Mamtaz & Bache 2001), ion-exchange (Polowczyl et al. 2017), and adsorption (Shan et al. 2016). Among these methods, adsorption has gained wide acceptance because of high efficiency, low cost and easy operation (Lou et al. 2015). However, most adsorption studies for Mo(VI) removal are carried out in an oxic environment, and there are few studies on Mo(VI) removal in anoxic water bodies. Iron monosulfide (FeS) is an important natural reductive mineral that exists extensively in soil, river sediment, underground water and coastal waters (Richard & Luther 2007). It plays important roles in the sorption, immobilization, and transformation of a variety of heavy metals and inorganic oxyanions in reducing environments, because of unique molecular structure and surface chemical properties (Morse & Arakaki 1993; Gong et al. 2016; An et al. 2017). Note that FeS has a very important influence on the geochemical cycle of molybdenum, because the content of soluble molybdenum in sulfide sediments, estuaries, and other reducing environments is extremely low (Crusius et al. 1996; Morgan et al. 2012). It is presumed that Mo(VI) may be reduced to MoS₂, or it may be associated with iron sulfide, or sulfide ferrous minerals have undergone synergistic precipitation (Bostick & Fendorf 2003). Nanoscale zerovalent iron (nZVI), as an environmental green material (Li et al. 2009), has a large surface area and is inexpensive, highly reducible, suitable for onsite operation, and can be magnetically separated (Cundy et al. 2008; Fan et al. 2009; Fu et al. 2014; Guan et al. 2015). The use of nZVI for removing heavy metals ions, such as Mo(VI), Pb(II), Cr(VI), and As(V), from contaminated water and soil has shown promising results (Yan et al. 2010; Boparai et al. 2011; Huang et al. 2013). It has been reported that nZVI can remove Mo(VI) by adsorption of Mo(VI) onto the surface of nZVI followed by reduction to Mo(IV) (Qian et al. 2018). However, Fe⁰ is easily passivated in water, rapidly lowering the rate of Mo(VI) reduction.

Recent studies show that sulfide-modified nZVI (S-nZVI) has higher reactivity, longevity and selectivity than nZVI and has great potential for the environmental remediation of heavy metals (Wu et al. 2018). It has been shown to increase precipitation and inhibit outer-sphere complexation for arsenic removal and to increase electron transfer from the nZVI core to the surface of Cr(VI) (Li et al. 2018; Wu et al. 2018). It can be expected that S-nZVI could be used for the removal of Mo(VI) from contaminated water. However, S-nZVI has not yet been studied for Mo(VI) removal; the specific mechanisms and the key influencing factors are not clear. Therefore, the goal of this study is to investigate S-nZVI for Mo(VI) removal from aqueous solution under anoxic conditions. Because S:Fe molar ratio plays an important role in the synthesis of S-nZVI, three S:Fe molar ratios (0.1, 0.5, and 1.0) on the structure of S-nZVI and the removal of Mo(VI) were examined. Also, we comprehensively analyzed, characterized and compared the composition, structure and morphology of S-nZVI before and after reaction with Mo(VI) to determine the mechanisms of Mo(VI) removal. To comprehensively evaluate Mo(VI) removal capacity by S-nZVI, the potential influences of pH, initial Mo(VI) concentration, S-nZVI dose, cation content, competing anions, ionic strength, and aging time were also studied. The obtained results would provide a theoretical reference for the treatment of Mo(VI) pollution in anoxic water bodies. It is crucial for the cognition of the migration and transformation process of Mo(VI) pollution in anoxic water bodies (such as groundwater and bottom water bodies of rivers and lakes).

Chemicals used in this study were analytical grade (see Text S1 in Supporting Information for more details). Experiments were conducted in a gloved anaerobic chamber with an atmosphere of high purity nitrogen (99.99%).

Both nZVI and S-nZVI were synthesized using the method described by Li et al. (2016), with minor modification. In each trial, 180 mL of 0.3 M NaBH₄ solution containing a variable amount of Na₂S₂O₄ was added to 60 mL of 0.2 M FeCl₂ solution to obtain a black precipitate of S-nZVI (S:Fe molar ratios = 0.1, 0.5, and 1.0) or nZVI. The precipitate was centrifuged at 8,000 rpm for 5 min, washed in deoxygenated deionized water (three times) and then used in the experiments.

All batch experiments were conducted anaerobically at constant temperature in the anaerobic chamber using 100 mL screw-capped vials as reactors, unless specified otherwise. The batch reactors were wrapped to keep them dark and mixed using a magnetic stirrer at 180 rpm. The samples were collected at regular time intervals in a 10 mL syringe and analyzed immediately after they were filtered through a 0.45 mm filter membrane. Batch experiments were conducted to examine the effects of four factors on the removal of Mo(VI) by S-nZVI: S:Fe ratio (0.1, 0.5 or 1.0); initial dose of S-nZVI (200, 500, 800 or 1,000 mg/L); and Mo(VI) concentration (20, 50, 80 or 100 mg/L). Such a concentration range of Mo(VI) has been found in various environments (Table S1). The enrichment of molybdenum in the iron sulfides environment is usually accompanied by the change of the solution pH (Bostick & Fendorf 2003). Therefore, the solution pH (4.0–9.0) was adjusted by buffer solutions (Bostick & Fendorf 2003) to obtain a constant ionic strength that allows for comparison between different experiments. The equilibrium pH was adjusted with different pH buffers: HAc–NaAc (pH 4–5), MOPS–NaAc (pH 6–7), H₃BO₃–Na₂B₄O₇ solution (pH 8–9). The experimental control conditions were constant: S:Fe ratio 0.5; pH 7.0; Mo(VI) concentration 50 mg/L; and S-nZVI dose 500 mg/L. When one of the four factors was investigated, the other three factors were kept at constant values. Note that control pH was 7.0 rather than 4.0 to examine the effect of initial Mo(VI) concentration and S-nZVI dose on Mo(VI) removal by S-nZVI because preliminary experiments showed that the reaction rate parameter of Mo(VI) at pH 4.0 was too rapid to be determined accurately. The results were analyzed using pseudo-first-order, pseudo-second-order, and Weber–Morris diffusion kinetics models to quantify Mo(VI) removal efficiency. Details of the models and the methodology of determining the parameters are provided in Supporting Information (Text S2).

Experiments were also conducted to investigate the influence of other factors on the removal of Mo(VI) by S-nZVI: co-existing cations (Zn²⁺, Ca²⁺, Mg²⁺, Cu²⁺, Ni²⁺ and Co²⁺, used as their respective 0.01 M chlorates); and competing anions (⁠⁠, in the form of a 0.5 mM sodium salt, basically consistent with Mo(VI) concentration, and as a 0.1 M sodium salt). The effect of ionic strength was tested in the presence of 0.05–0.5 mol/L KCl or without background electrolyte. These experiments were conducted under experimental conditions similar to those described above. Both 500 mg/L S-nZVI particles and 100 mL deoxygenated deionized water were introduced into the 100 mL screw-capped vials to investigate the effect of aging time on removal of Mo(VI) by S-nZVI. To simulate the environment of natural anoxic water bodies, the vials were sealed and placed in a dark room for aging. The aged samples were collected after various time intervals (0, 1, 7, 14, 21, 28 or 35 d) and used for Mo(VI) removal following the procedure described above.

Separate experiments were conducted to investigate changes in surface morphology and structure of S-nZVI before and after reaction with Mo(VI). In each trial, after 1 h reaction with Mo(VI), the S-nZVI suspension was centrifuged at 8,000 rpm for 5 min. The S-nZVI solids were dried under nitrogen and used for surface analysis. Concentrations of Mo(VI) and total Fe ions were measured by flame analysis using a polarized Zeeman atomic absorption spectrophotometer (Hitachi ZA3000). All experiments were conducted in triplicate, and the results presented as mean value ± standard deviation. All graphs were drawn using Origin 8.5. Details of the analytical methods for surface characterization of the synthetic S-nZVI are given in Text S3 in Supporting Information.

S-nZVI with different S:Fe molar ratios (0, 0.1, 0.5 or 1.0) was obtained by adding different concentrations of Na₂S₂O₄ during nZVI synthesis. As shown in Figure 1(a), the performance of S-nZVI in removing Mo(VI) is strongly influenced by the S:Fe molar ratio. When the S:Fe molar ratio was initially increased from 0 to 0.5, Mo(VI) removal efficiency increased from 12.46% to 88.73%. That is, the Mo(VI) removal efficiency of S-nZVI with S:Fe ratio 0.5 was seven times greater than that of nZVI. This increase in Mo(VI) removal by S-nZVI is explained as follows. Sulfur binds with Mo(VI) to increase the adsorption of Mo(VI) on the surface of S-nZVI; increased sulfur on the surface creates more binding sites for molybdenum (Helz et al. 2004). However, as the S:Fe molar ratio increased from 0.5 to 1.0, Mo(VI) removal efficiency decreased from 88.73% to 69.16%, indicating that higher sulfur content decreased Mo(VI) removal efficiency of S-nZVI. The quantity of released Fe in the final supernatant after Mo(VI) reaction with S-nZVI was determined for different S:Fe ratios. The results showed that dissolved Fe in the solution after Mo(VI) was reacted with S-nZVI was at a lower concentration than when Mo(VI) was reacted with nZVI (Figure S1), indicating that sulfidation inhibited corrosion of nZVI. Moreover, the FeS layer may be an electron conductor that accelerated the multi-electron transfer from the Fe⁰ core to Mo(VI) oxyanions, as described by Du et al. (2016). Obviously, S-nZVI with an S:Fe ratio of 0.5 (S-nZVI_0.5) was clearly most effective in Mo(VI) removal. Other studies have also found that S-nZVI_0.5 has very high reactivity with tetrabromobisphenol A and hexabromocyclododecane (Li et al. 2016, 2017).

The effects of initial dose on Mo(VI) removal by nZVI and S-nZVI_0.5 at pH 7.0 are shown in Figure 1(b) and 1(c). The results show that, at the same dose, nZVI is much less effective than S-nZVI_0.5 in removing Mo(VI). Also, Mo(VI) removal efficiency increased rapidly from 43.79% to 98.20% within 60 min of the reaction beginning as S-nZVI dose increased from 200 to 1,000 mg/L (Figure 1(c)). The explanation for this is that the surface area and number of reaction sites both increased on S-nZVI particles, thus improving the removal of Mo(VI). When the initial concentration of increased from 20 to 100 mg/L, there was a significant decrease in Mo(VI) removal efficiency from 97.41% to 37.85% within 60 min of the reaction beginning (Figure 1(d)). The number of available S-nZVI adsorption sites did not change, but a greater initial concentration of requires more adsorption sites for complete removal; if there are insufficient sites available, the removal rate will decrease for a higher initial concentration.

To identify the surface reaction mechanisms of S-nZVI with Mo(VI), the surface of S-nZVI was investigated. Brunauer–Emmett–Teller specific surface area (BET-SSA) analysis showed that the specific surface areas of S-nZVI particles with S:Fe ratios of 0.1–1.0 were about 30–40 times larger than those of nZVI (Table S2). Moreover, the specific surface area of S-nZVI_0.5 was larger than that of S-nZVI_1.0, which helps to explain the Mo(VI) removal efficiency decrease as S:Fe molar ratio increased from 0.5 to 1.0. Transmission electron microscopy (TEM) showed that the structure of unmodified nZVI appeared to be chains of spherical particles with smooth surfaces, while S-nZVI_0.5 particles had a core of nZVI surrounded by a flake-like shell (Figure 2). After reaction with Mo(VI), the core–shell structure of S-nZVI_0.5 disappeared. A rough structure and flocculent precipitate formed on the surface, likely due to molybdenum binding and oxidation of divalent iron ions (Sun et al. 2017). Energy-dispersive X-ray spectroscopy (EDS) confirmed the presence of Fe, S, and Mo on the surface of S-nZVI after reacting with Mo(VI) (Figure S2). The elemental maps of Fe, S, and Mo suggested that these elements were well dispersed, which is consistent with EDS analysis.

The preceding solution and solid phase analyses suggest that there are three main stages of Mo(VI) removal: aqueous Mo(VI) oxyanions diffused to the water–S-nZVI interface, a process that is driven by valence forces and electrostatic attraction between sulfide ions, iron ions and Mo(VI) oxyanions; Mo(VI) was adsorbed at the solid–liquid interface by co-precipitation; and Mo(VI) was simultaneously reduced to Mo(V). Thus, the removal of Mo(VI) using Fe⁰-based materials is a multi-step process that involves adsorption as well as a subsequent Mo(VI) reduction (Huang et al. 2012; Qian et al. 2018). To describe the Mo(VI) removal route by S-nZVI_0.5, pseudo-first-order and pseudo-second-order kinetics models were used to represent the reaction (Figure S3 and Table 1). It was observed that the experimental data fitted well with the pseudo-second-order adsorption model, with high correlation coefficients (R² > 0.996). The adsorption capacity obtained by the experiment (61.25 mg/g) was found to be close to the theoretical values (65.36 mg/g; reaction conditions: Mo(VI) concentration 50 mg/L, S-nZVI_0.5 concentration 500 mg/L, pH 7.0) predicted by the pseudo-second-order kinetic model. These data show that the pseudo-second-order adsorption model best represented the adsorption process, which implies that chemisorption was the rate-limiting process for Mo(VI) removal.

Solute adsorption onto the solid surface can be controlled by several steps, e.g., external diffusion, surface diffusion, and pore diffusion (Debnath & Ghosh 2009). In order to confirm the actual rate-controlling step in the Mo(VI) adsorption process, the well-known Weber–Morris equation was applied (Weber & Morris 1963). The plots of q_t versus t^0.5 are given in Figure S4 for the adsorption of Mo(VI) by S-nZVI_0.5 at different initial Mo(VI) concentrations and S-nZVI_0.5 dose. Obviously, the adsorption data were well-fitted by two straight lines, indicating that two steps took place during Mo(VI) adsorption onto S-nZVI_0.5. The detailed fitting model parameters are shown in Table 2. The R² values are close to unity, also confirming the applicability of the Weber–Morris model for Mo(VI) adsorption on S-nZVI_0.5. Firstly, Mo(VI) in aqueous solution was transported onto the surface of S-nZVI_0.5, then Mo(VI) was transported and adsorbed on the interior surface of S-nZVI_0.5 (Mittal et al. 2007). Also, the time in intraparticle diffusion state is longer than that of external diffusion state, indicating that the intraparticle diffusion is the main rate-controlling state.

The initial solution pH was changed from 4.0 to 9.0 to investigate the effect of pH on removal by S-nZVI. As shown in Figure 4(a), after 60 min, removal efficiency of Mo(VI) was 3.3%, 10.1%, 78.0%, 89.6%, 93.8% and 91.6% at pH of 9.0, 8.0, 7.0, 6.0, 5.0 and 4.0, respectively, suggesting that Mo(VI) removal increased with the decrease in solution pH. Preferential removal of Mo(VI) in an acidic solution is consistent with previous adsorption results (Lian et al. 2019). Acidity usually affects adsorption by changing the form of Mo(VI) and altering the surface zeta potential of adsorbents. The predominating ionic species of Mo(VI) were obtained from the Visual MINTEQ 3.0 software calculation. Obviously, at a concentration of 50 mg/L can be changed to other species that are more favorable for adsorption by electrostatic interaction at pH 4.0–7.0, such as ⁠, or (Figure 4(c)). Zeta potential of S-nZVI was measured as a function of pH to determine the surface charge. The pH of point of zero change (pHpzc) of S-nZVI was calculated to be around 7.6 (Figure 4(b)). Thus, the positive surface charge attracted the negatively charged Mo(VI) species when pH was <7.6. When pH increased to be above pHpzc, electrostatic repulsion between the negatively charged surface of S-nZVI and negatively charged Mo(VI) species inhibited the adsorption of Mo(VI) and decreased removal efficiency. Nevertheless, the removal efficiency of Mo(VI) at pH = 4.0 was lower than that at pH = 5.0, which may be due to the change in the iron sulfides species. Richard & Luther (2007) reported that FeS can react with H⁺ to form Fe²⁺ and HS⁻ under very acid condition. The loss of FeS on the surface of S-nZVI may contribute to the decrease in the capability of S-nZVI towards Mo(VI) removal.

Ionic strength affects binding of Mo(VI) with adsorbates by influencing the thickness and interface potential of the double electron layer of the adsorbents (Sari et al. 2007). Ionic strength can be used to distinguish between outer-sphere complexation and inner-sphere complexation. The influence of ionic strength on immobilization of Mo(VI) onto S-nZVI is shown in Figure 5(a). Increase in KCl concentration from 0 to 0.1 mol/L resulted in a decrease of 2.28 percentage points (from 67.58% to 65.30%) in the Mo(VI) removal rate. Outer-sphere complexation was inferred to be the removal mechanism of Mo(VI) because outer-sphere surface complexes are more susceptible to change in ionic strength than inner-sphere surface complexes (Kim et al. 2013). One reason for the decrease in Mo(VI) removal rate is the high concentration of Cl⁻, which competed with for the positively charged adsorption sites on S-nZVI (Elwakeel et al. 2009).

The presence of anions (⁠ and ⁠) also decreased the Mo(VI) removal rate (Figure 5(b)) because S-nZVI was positively charged during the reaction (as shown in Figure 4(b)) and electrostatically attracted the negative anions. Thus the two oxyacid ions competed for adsorption of on S-nZVI. However, a high concentration of reduced the inhibitory effect of the reaction, which was different from the effect of ⁠. Ferrous ions are oxidized by Cr(VI) at the solid–solution interface, but not in aqueous solution, and the thickness of the iron oxide shells can be a barrier for electron transfer (Du et al. 2016). Divalent or trivalent iron on the adsorbent surface showed good adsorption and complexation of Mo(VI) (Lian et al. 2019; Lian et al. 2020). Figure 5(c) shows the effect of co-existing cations on the removal of Mo(VI) from aqueous solution onto S-nZVI. Co-existing cations present in natural water (such as Zn²⁺, Ca²⁺ or Mg²⁺) increase Mo(VI) removal. An explanation of this effect is that the standard reduction potentials of Mg²⁺, Zn²⁺ and Ca²⁺ (E₀ = −2.37, −0.76 and −2.87 V, respectively) are less than that of Fe²⁺ (E₀ = −0.45 V), and thus any of the three free cations may form some covalent compound with Mo(VI), such as CaMoO₄. The specific chemical reaction needs further study. Cu²⁺, Co²⁺ and Ni²⁺ that usually present in industrial wastewater have inhibitory effect on the Mo(VI) removal by S-nZVI. Cu²⁺, Co²⁺ and Ni²⁺ (E₀ = +0.34, −0.28 and −0.26 V, respectively) have greater positive standard reduction potentials than Fe²⁺ and may occupy more active sites on the surface of S-nZVI. In other words, they may result in insufficient adsorption sites for Mo(VI) reactions.

The removal efficiency of Mo(VI) of fresh nZVI and S-nZVI was 12.46% and 85.49%, respectively. After S-nZVI particles were aged in water for 35 days, the removal efficiency of Mo(VI) was 57.39%, which is still higher than that of fresh nZVI. It suggested that the aged S-nZVI had a higher capacity for Mo(VI) removal than freshly-prepared nZVI (Figure 6). It is consistent with the results of previous studies that S-nZVI has higher longevity than nZVI (Li et al. 2016). It is generally accepted that sulfidation of nZVI can inhibit the H₂ evolution reaction (reduction of water by Fe⁰ to form hydrogen), reduce the Fe⁰ corrosion (Chin et al. 2005; Liu & Lowry 2006), decrease the magnetic force of aged S-nZVI (Kim et al. 2014), and therefore improve the long-term performance of nZVI in the removal of contaminant. Overall, S-nZVI is more stable and long-lasting than nZVI in its capacity to remove Mo(VI).

This is the first report on the reactivity of S-nZVI towards Mo(VI) under various aquatic chemistry conditions. S-nZVI had higher reactivity towards Mo(VI) than did nZVI. Surface analysis indicated that Mo(VI) could be removed by adsorption, co-precipitation, and reduction. The Mo(VI) removal by S-nZVI is well described with the pseudo-second-order adsorption model. Batch experiments showed that S-nZVI removed Mo(VI) at faster rate under acidic to neutral pH (4.0–7.0) than alkaline condition. The ions ⁠, ⁠, Ni²⁺, Cu²⁺ and Co²⁺ inhibited the removal of Mo(VI) whereas Zn²⁺, Ca²⁺ and Mg²⁺ enhanced it. Aging experiments demonstrated that S-nZVI after aging of 35 days in water still had greater Mo(VI) removal capacity compared to nZVI, indicating the feasibility of S-nZVI for treatment of Mo(VI)-contaminated waters under anoxic condition.

This work was financially supported by the National Natural Science Foundation of China (Nos. 51709001, 41773132 and 42077285), Guangdong Foundation for Program of Science and Technology Research (Nos. 2017B030314057, 2020B1212060053 and 2019B121205006), the State Key Laboratory of Organic Geochemistry, GIGCAS (No. SKLOG2020-4), and the China Scholarship Council for Jianjun Lian (No. 201808340035).

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