Novel Styrene-Based Polyamine Sorbent for Efficient Selective Separation of Molybdenum.

Tungsten (W) and molybdenum (Mo) are important strategic resources but the two coexist in both primary ore and waste. Before a single metal product is obtained, it is often necessary to separate the two. In this work, we reported two new polyamine resins (D301@PA and D301@TA), which can be obtained by an assembled amine (primary amine or tertiary amine) and traditional D301 resin by the dipping method. Then, the sorption experiments with the amine resins were carried out, and the selectivity and sorption capacity of the two new polyamine resins for MoS4 2- have been significantly improved. Among them, D301@TA showed the highest sorption capacity of 414 mg·g-1 and a separation factor of 108. Finally, the sorption mechanism can be inferred through scanning electron microscopy (SEM), Brunauer-Emmett-Teller (BET) analysis, and X-ray photoemission spectroscopy (XPS); the Cl- ions in the amine resin and the MoS4 2- ions were subjected to ion exchange. This work provides a green and efficient approach for separating tungsten and molybdenum.

Introduction

With the increase of the global population and the continuous economic growth of developing countries, the demand for energy is increasing worldwide, especially the refractory metals tungsten and molybdenum required in production and life. Recently, increasing attention has been paid to the recycling of the secondary resources of tungsten and molybdenum.[1] The recovery of tungsten and molybdenum secondary resources can not only reduce the environmental pollution caused by stacking but also effectively supplement the tungsten or molybdenum mineral resources and reduce the damage to the environment during the mining of mineral resources.[2] Typical secondary resources of tungsten include tungsten slag, alloy scraps,[3] and waste catalysts containing tungsten,[4] but they are often accompanied by molybdenum because of their similar properties. Due to the purity requirements of tungsten products, the separation of tungsten and molybdenum has become a necessary part in the production process.[5,6] In particular, the industrial feed solution contains 30–160 g·L–1 WO3 and 1–3 g·L–1 Mo, which has a serious negative impact on the quality of tungsten products and causes the waste of molybdenum resources.[5]

To date, many techniques have been used to selectively remove molybdenum from tungstate solutions, which mainly include wet separation[6−8] and solid-phase sorption.[9] Traditional tungsten and molybdenum secondary resource recovery technology is based on solvent extraction technology,[7,10] which involves large material and energy consumption and produces organic waste liquid, waste gas, and other secondary pollution. Obviously, solid-phase sorption[11,12] (porous resin and carbon material) is simple, reliable, and efficient. So far, various porous materials such as porous resins, activated carbon, etc., used to adsorb molybdenum have been studied. The porous resin formed by the strong interaction between a macroporous weakly basic resin matrix and a nitrogen-containing functional group is the most common sorbent in tungsten production.[13] It is a promising sustainable separation technology for tungsten and molybdenum.

The application of styrene-based resins for separation of tungsten and molybdenum is low cost and environmentally friendly. To further improve the selectivity and sorption capacity of resins, a series of resins with styrene-based functional groups through chemical modification, impregnation, and other different methods were prepared. Due to the excellent affinity between N and C, the amine functional group can be assembled on the styrene matrix through a chemical reaction. For example, Cr has been recovered with a poly-epichlorohydrin-dimethylamine-modified (EPIDMA-modified) D301 resin.[14] Thus, to improve the absorption capacity, the functional group could be modified.[15,16] These experiments show that the resin’s selective sorption to metal ions is mainly caused by the amino group in the resin; the longer the carbon chain contained in the amino group, the greater the steric hindrance, resulting in better selectivity. However, the number of sorption sites that can be provided by each single resin molecule is limited, and the amino group sneaks into the depth of the resin’s pore size and cannot effectively contact the molybdenum molecule. Therefore, it is urgent to develop a novel styrene-based polyamine sorbent for efficient selective separation of molybdenum. Recently, primary amines (PAs) and tertiary amines (TAs) have attracted enormous attention owing to their excellent properties used for separation of tungsten and molubdenum.[6,17]

In this study, to solve the above-mentioned shortcomings of the styrene matrix and to further explore and study the sorption performance and influence of different amino groups on MoS42– anions, we prepared two kinds of new polyamine sorption materials (D301@TA and D301@PA). The aims of separating tungsten and molybdenum can effectively be achieved. Sorption–desorption experiments confirmed the sustainability and cost-effectiveness. In comparison to other similar adsorbents, the resin has advantages of easy preparation, low cost, high sorption capacity, and environmentally friendly. Finally, the sorption mechanism was thoroughly analyzed through SEM, BET analysis, and XPS.

Results and Discussion

Sorption Properties

To evaluate the sorption performances of D301, D301@PA, and D301@TA for MoS42–, a series of experiments were performed under different conditions, and the results are shown in Figure . The relevant experimental data are shown in Tables S1–S4 in the Supporting Information. First, the pH value of the tungstate solution was adjusted by adding diluted hydrochloric acid or sodium hydroxide. As shown in Figure a, D301, D301@PA, and D301@TA can remove MoS42– effectively in a wide pH range (from 5.2 to 10.2). QMo increased with the increase of the pH value until it exceeded 7.2. Then, upon further increasing the pH value, the QMo value decreased because H+ ions in the solution saturated MoS42– anions under acidic conditions, thereby reducing the sorption capacity of the adsorbent. As the pH value increased, the concentration of OH– in the solution increased. Since the electron-donating ability of OH– is stronger than that of MoS42–, N atoms in the resin will be preferentially saturated by OH– rather than MoS42–. Thus, a pH value of 7.2 was selected as the working condition. The reaction temperature also changed the sorption capacity. As shown in Figure b, with an increase in the temperature from 25 to 75 °C, the sorption capacity of resins for Mo slowly decreased. Considering both the sorption capacity and the separation factor, a reaction temperature of 25 °C might be appropriate for the sorption process.

Figure 1

Effect of the (a) pH value, (b) reaction temperature, (c) contact time, and (d) raw concentration of Mo in the solution on sorption capacity (Qe) and separation factor (β).

As shown in Figure c, the sorption capacities and separation factors of D301, D301@PA, and D301@TA increased as the contact time increased at 25 °C. It was observed that the sorption capacity increased for MoS42– as the reaction time increased from 0 to 150 min and then became constant. In the follow-up mechanism analysis, saturation of the active sites (N) in resins and electron transfer between adsorbents and MoS42– anions occurred. The active sites (N) were dominant for 60 min, so the sorption capacity increased rapidly. After the reaction of active sites was completed, the adsorbent continued to play its role, and the quality of the adsorbent was still increasing slowly until it reached an equilibrium. As shown in Figure d, we changed the mass of Mo in the raw solution to simulate different Mo concentration environments and controlled the reaction temperature at 25 °C; the contact time was 4 h. The results showed that the maximum sorption capacities of MoS42– by D301, D301@PA, and D301@TA were 157, 224, and 414 mg·g–1, and the separation factors were 45, 71, and 108, respectively.

As shown in Figure , when PA or TA was loaded on D301, its sorption capacity to MoS42– became higher than that of D301, especially in the case of TA, which is 257 mg·g–1 higher than that of D301, indicating that the proposed amine resin exhibited high sorption capacity and high selectivity to MoS42–. Finally, as shown in Figure , we summarized some typical molybdenum adsorbents[18] and several typical styrene-based adsorbents.[19−22] Obviously, compared with other molybdenum adsorbents materials, styrene-based materials have unique advantages in terms of MoS42– sorption capacity. Among the styrene-based materials, our reported D301 resins showed the best high capacity. Moreover, as purely organic solid materials, they are more conducive to the subsequent long-term storage, and their production cost is lower.

Figure 2

Comparison of the sorption capacity of some typical sorbents.

Structures

To study the adsorption mechanism of the adsorbents, the structures of the adsorbents were determined by solid nuclear magnetism. Figure shows the solid-state 13C results of D301, D301@PA, and D301@TA. The peak position shows a little change because the relative content of the N-containing functional group is small, which can still explain the changes in the molecular structures before and after modification. In the solid phase, the characteristic peak at 128.5 ppm can be assigned to the benzene ring of the resin. In addition, the peaks at 145 ppm are attributed to the carbons (−CH2CH3−) in the benzene ring. These were observed in the resin before and after modification, indicating that the benzene ring structure is stable. The peaks at 43 and 65.5 ppm can be assigned to the functional groups containing N. Based on this, the molecular structural formulas of the three resins can be inferred. The different sorption capacities of the three resins for molybdenum in the solution drawn from Figure may be mainly caused by the different structures of their N-containing functional groups. In summary, the above results show that the sorption performance of the resins is closely related to their molecular structures.

Figure 3

13C spectra of (a) D301, (b) D301@PA, and (c) D301@TA.

To further study the relationship between the molecular structure and sorption performance, first-principles calculations were employed to explore the relationship at a nanoscale. For these calculations, density functional theory (DFT) was used,[23−25] and the molecular structures were based on the results of NMR analysis. To establish the relationship between the molecular structure and analytical properties, it is necessary to perform optimization calculations on the structure. As shown in Figure , the optimized structures of D301, D301@PA, and D301@TA were obtained under the generalized gradient approximation (GGA), and the exchange–correlation functional of the calculation process uses the PBE functional. The resins, well known as anion exchange materials, have more active highest occupied molecular orbitals (HOMOs). According to frontier orbital theory, the smaller the absolute value of the energy difference between the HOMO of one reactant and the lowest unoccupied molecular orbital (LUMO) of another reactant (ΔE), the more favorable is the interaction between the two. The efficiency of the amine resin as an adsorbent for MoS42– (donor and back-donor) depends on the ability to donate electrons and the tendency to accept them. Accordingly, it becomes necessary to discuss ΔE. The reaction energy of the resin with MoS42– was calculated according to eq , and the results are shown in Table .where EHOMOresin refers to the HOMO energy of the resin of the model compound, and ELUMOMoS refers to the LUMO energy of the model compound of MoS42–.

Figure 4

Optimized structures of (a) D301, (b) D301@PA, and (c) D301@TA obtained by MS DMol3.

Table 1

ΔE Data Based on Frontier Orbital Theory

model	LUMO (Ha)	HOMO (Ha)	ΔE (Ha)
D301 resin	–0.639	–4.407	4.300
PA resin	–0.338	–0.919	0.812
TA resin	–0.166	–0.326	0.219
MoS42–	–0.107	–0.180

According to Table , the results show that the ΔE values (0.812 and 0.219 Ha in solution) are relatively small, especially that of D301@TA, illustrating that it is easier to adsorb molybdenum in solution. Both the above trend of ΔE and the HOMO orbital value are consistent with that of the sorption performances in Figure , indicating that the lower energy may be one of the reasons for the excellent selectivity of resin molecules to Mo. In conclusion, different resin molecules form different morphologies, and the intrinsic energy values are determined by the different molecular structures, resulting in different adsorption properties. At the same time, this work realized the modification of D301 resin, the adsorption capacity, and the ability to donate electrons; D301@TA was better than D301@PA.

Morphology

To further understand the changes in the morphology of the resin before and after modification, Figure shows the morphologies of D301, D301@PA, and D301@TA. For D301, from Figure a, it can be clearly seen that the resin is composed of macropores and columnar cross-linked pores. After modification, the macropores disappeared significantly. For D301@PA, as shown in Figure b, almost all of the macropores disappeared and a more aggregated pore structure appeared. It can be seen more clearly in Figure c that the resin surface of D301@TA after TOA modification becomes a regular pore structure. Compared with D301, the surface morphology and pore structure of D301@PA and D301@TA are more regular, especially those of D301@TA, which is more conducive to sorption of MoS42– ions with a radius of 0.330 nm, and that can well explain the excellent sorption capacity of D301@TA.

Figure 5

SEM images of (a) D301, (b) D301@PA, and (c) D301@TA.

Specific Surface Area and Pore Size Distribution

The specific surface area and pore size distribution of D301, D301@PA, and D301@TA were analyzed by the N2 adsorption–desorption isotherm and the BJH model, and the results are shown in Figure . According to the calculations of BET and BJH, the specific surface area of pure D301 is 26.735 m2·g–1 and the pore size distribution is mainly concentrated at 17.510 nm. However, when the surface of pure D301 was coated with PA and TA aminating agents, the specific surface areas of the two modified adsorbents decreased to different degrees, corresponding to 25.341 and 22.981 m2·g–1, respectively, and the pore size distribution also changed. The reasons for the different pore size distributions are presented as follows: First, combined with the SEM analysis mentioned before, it can be more easily understood that compared with pure D301, the modified D301 surface was covered with a layer of the aminating agent, which blocked the original pores and channels, thereby reducing the specific surface area and pores. Second, PA and TA have stretchable carbon chains and they can combine with D301 through bonding during the preparation process, and the carbon chains are bent to increase the pore size distributions of D301@PA and D301@TA (Figure d–f). As shown in Figure d–f, it is obvious that all three are in the range of mesopores (2–50 nm), and their pore size is significantly larger than the radius of the MoS42– anion (0.330 nm). Therefore, the sorption of the resin to MoS42– is not only the surface coverage of the sorbent material, and the sorption capacity of the sorbent cannot be described by the specific surface size alone. Actually, the sorption capacity of the resins to MoS42– anions is mainly concentrated in the filling of micropores (below 2 nm). The increase in the pore size distribution promotes faster passage to MoS42– molecules with a larger radius during the sorption process, so as to accelerate the sorption reaction between the adsorbents and MoS42– molecules.

Figure 6

N2 adsorption–desorption isotherm (a–c) and corresponding pore size distributions (d–f) of (a, d) D301, (b, e) D301@PA, and (c, f) D301@TA.

Ion Exchange

The XPS full spectrum analysis of the adsorbent before and after sorption is performed. Figure a shows the XPS full spectrum results of the three adsorbents before adsorption, and Figure b shows the element distribution on the resin surface after adsorption. The resin is mainly composed of C, N, and Cl elements before adsorption. After adsorption, the Cl element disappeared and the distribution of Mo and S elements appeared. It is speculated that the resin undergoes anion exchange before and after adsorption, and the Cl– anion on the resin was replaced by MoS42– in the solution species.

Figure 7

XPS spectra of D301, D301@PA, and D301@TA (a) before and (b) after sorption of MoS42–.

Sorption Bonding

To further analyze the changes before and after the sorption of D301, D301@PA, and D301@TA, Figure shows a typical XPS N atom analysis. Among them, Figure a shows the N spectrum results of the three adsorbents before adsorption, and Figure b shows the N spectrum after adsorption. Through the comparison of N spectra before and after adsorption, the bonding relationship of the atoms can be accurately analyzed. First, Figure a shows the curves of D301, D301@PA, and D301@TA, where 402 and 399 eV are attributed to N+. Interestingly, compared with D301, the binding energies of D301@PA and D301@TA increased by 0.5 eV, which may be caused by their lattice defects. A quarter of the nitrogen ion positions in the crystal lattice are empty and have very high nitrogen ion conductivity, which leads to an increase in binding energy. After sorption, blue shifts of C–N bonds (from 402 to 401.6 eV and from 399 to 398.8 eV) are observed. Blue shifts can reduce the electron density around the N atoms, which makes it easier to obtain electrons to form N–MoS42– bonds. Fu et al.[26] reported that high-resolution XPS spectra of the nitrogen element reveal the blue shift of the pyridine N 1s peak (from 398.6 to 399.0 eV) and the ternary amine N 1s peak (from 400.4 to 401.0 eV) due to the stronger interaction between dipicolylamine groups and MoS42– ions.

Figure 8

N spectra of D301, D301@PA, and D301@TA before (a) and after (b) sorption of MoS42–.

Reusability

Compared with other sorption materials, polymer materials generally possess relatively high regeneration capacity in acid or alkaline elution. A 5 wt % sodium hydroxide solution was used as the eluent; the sorption data after five cycles are shown in Figure . It can be seen from Figure that the sorption performance of the adsorbent has not changed significantly; therefore, the resin is cost-effective. Considering the sorption capacity and sustainability, we infer that D301@TA has the best effect in large-scale industrial use.

Figure 9

Regenerative performance of D301, D301@PA, and D301@TA at pH = 7.2 (contact time: 240 min and reaction temperature: 25 °C).

Stability

To verify the stability of the resin, the morphology of the resin after five cycles of use was characterized, as shown in Figure . As can be seen, after five cycles, the pore structures of the modified resins ((b) D301@PA and (c) D301@TA) were obviously more regular than that of the resin before modification, which were also present before sorption. Moreover, compared with the resin before sorption, the pore structure has not changed. It is speculated that the stability of this pore structure is one of the main reasons for the resin to maintain its reusability.

Figure 10

SEM images of (a) D301, (b) D301@PA, and (c) D301@TA after sorption of MoS42–.

Conclusions

In this study, two styrene-based sorbents, D301@PA and D301@TA, were synthesized by the hydrothermal method. Both sorbents use the styrene matrix as the skeleton, and different aminating agents N1923 and TOA are used as modifiers. They are uniformly modified on the surface and inside to increase the specific surface and pore size distribution, thereby increasing the sorption capacity of the resin. The experimental results show that the saturated sorption capacities of D301@PA and D301@TA reach 224 and 414 mg·g–1 at 25 °C for 4 h. Moreover, by XPS characterization, the sorption mechanism of these two materials was deeply revealed. The whole sorption process can be divided into two steps. The first step was the ion exchange between the Cl– anion on the resin and MoS42– in the solution. Then, the S atoms in MoS42– and the N atoms on the resin formed N–S bonds through electron transfer, and MoS42– anions were firmly adsorbed on the resins. By adjusting the experimental and characterization results, the modified resins D301@PA and D301@TA can be obtained and their sorption capacities to Mo were better than that of commercial D301 resin. Finally, further in-depth research, especially the reusable sorption–desorption experiments under working conditions, demonstrates that graft modification may be an effective method to further improve the capacity of the resin for practical industrial applications.

Experimental Section

Materials

The macroporous styrene matrix (D301) used in this study was purchased from Shanghai Kaiping Chemical Co., Ltd., China. N,N-Dimethylformamide (DMF, purity ≥ 99%), primary amine N1923 (PA, purity ≥ 95%), trioctylamine (TA, purity ≥ 95%), Na2WO4·2H2O (purity ≥ 99.5%), and Na2MoO4·2H2O (purity ≥ 99.5%) were kindly provided by Shanghai Macklin Biochemistry Co., Ltd., China. All of the other chemicals in this work were of AR grade. The stock solution containing tungsten and molybdenum was prepared by dissolving 66.4 g of Na2WO4·2H2O and 2.42 g of Na2MoO4·2H2O (both purity ≥ 99.9%) in 1 L of deionized water. Then, the solution was vulcanized ([S]/[Mo] = 6) for 2.5 h at 72 °C.

Sorbent Preparation

First, D301 resin was washed with distilled water until a neutral pH value was reached. Then, the resin was swelled in DMF (mass/vol = 1:4) and aminated in PA or TA (mass/vol = 1:4) to prepare D301@PA and D301@TA. The simple preparation procedure of two adsorbents is shown in Figure .

Figure 11

Schematic diagram for the preparation of D301@PA and D301@TA adsorbents and the MoS42– sorption experiment.

Sorption Experiment

Sorption experiments were performed by shaking 20 mL of raw solution with 0.02 g of amine resin in a 30 mL Erlenmeyer flask. To reduce the experimental error, each group of experiments was conducted three times. Then, the metal concentration was determined by the standard deviation of the triplex analyzed values. The effects of pH values (5.8–11.8), reaction temperatures (25–75 °C), and reaction times (1–6 h) on the sorption capacity of the amine resin were studied. Then, the aqueous phase was separated to determine the concentration of the metal elements. The concentrations of W and Mo in the amine resin were calculated according to the rule of mass balance. The sorption capacity (Qe), distribution ratio (D), and separation factor (β) were calculated according towhere C0 (g·L–1) is the raw concentration of Mo in the solution, and Ce (g·L–1) is the Mo concentration after sorption equilibrium.

Characterization

A Thermo Scientific iCAP 6000 series inductively coupled plasma optical emission spectrometer (ICP-OES, PE) was used to confirm the concentration of metals in aqueous solution. Detailed information on the chemical states and electronic structures of Mo, S, N, C, and Cl on the surface was obtained via X-ray photoemission spectroscopy (XPS, Thermo Fisher Scientific). The N 1s spectra were measured using a KRATOS Axis Ultra X-ray photoelectron spectrometer equipped with a monochromatic Al Kα source and a charge compensation system. During data collection, the step size was set to 0.05 eV, and 10 scans for each sample were averaged. All NMR spectra were recorded on a Bruker Inc., Switzerland (13C: 400 MHz). Scanning electron microscopy (SEM) was performed on a Gemini SEM 300 (Germany) to study the morphologies of the three materials before and after modification. The N2 adsorption–desorption isotherms of the three materials were tested using an ASAP 2020 apparatus utilizing BET and BJH models (Quantachrome Instruments).

DFT Calculations

Density functional theory (DFT) is a quantum mechanical method for studying the distribution of electrons in multielectron systems.[25,27,28] It uses electron density instead of wave functions as the basic quantity for the study, simplifying the interaction between electrons in multielectron systems. Some scholars have used the DFT method to calculate the ground-state electronic structures of resin molecules.[29,30] In this study, the DMol3 module in Materials Studio (Accelrys) was used for calculation, and the exchange–correlation functional of the calculation process selects the PBE functional under the generalized gradient approximation (GGA).