Jong Hyuk Jeona,
Ana Belen Cueva Solaab,
Jin-Young Leeab,
Janardhan Reddy Koduruc and
Rajesh Kumar Jyothi
*ab
aMineral Resources Division, Korea Institute of Geoscience and Mineral Resources (KIGAM), Daejeon 34132, Korea. E-mail: rkumarphd@kigam.re.kr; Fax: +82-42-868-3421; Tel: +82-42-868-3313
bDepartment of Resources Engineering, Korea University of Science and Technology (UST), Daejeon 34113, Korea
cDepartment of Environmental Engineering, Kwangwoon University, Nowon-gu, Seoul 01897, Korea
First published on 27th January 2022
Vanadium and tungsten ion adsorption and desorption characteristics and separation conditions were investigated using a simple porous anion-exchange resin. Initially, systematic experimental research was performed using synthetic aqueous vanadium and tungsten solutions. To evaluate the vanadium and tungsten (50–500 mg L−1) isotherm parameters, adsorption was performed at pH 7.0 using 0.5 g of ion-exchange resin at 303 K for 24 h. Well-known adsorption models such as Langmuir, Freundlich, and Temkin were used. Vanadium was desorbed from the resin using HCl and NaOH solutions. In contrast, tungsten was not desorbed by the HCl solution, which enabled the separation of the two ions. The desorption reaction reached equilibrium within 30 min of its start, yielding over 90% desorption. We investigated the adsorption mechanism and resin stability with the aid of spectroscopic and microscopic analysis, as well as adsorption results. The applicability and feasibility of the resin was tested via recovery of both metals from real spent catalysts. The applicability and reusability results indicated that the resin can be used for more than five cycles with an efficacy of over 90%.
Metallic elements such as vanadium and tungsten have been recovered from spent catalysts via techniques including solvent extraction (liquid–liquid extraction), ion exchange, and precipitation. Luo et al. (2003) reported that 86.96% of ammonium metavanadate and 85 to 90 wt% ammonium tungstate were recovered from tungsten alloy scrap using a precipitation technique.1 The precipitation method is a simple and quick operation in which metal ions precipitate from a solution via the addition of a precipitant in the form of an aqueous or non-aqueous solution. However, since various elements are precipitated together in this process, treatment of impurities is difficult. This adversely affects the performance and efficiency of the product, and thus limits application of this technique. With regard to the solvent extraction method, acidic extracting agents such as D2EHPA, PC-88A, and Aliquat 336, and basic extractants including Alamine 336, are the most widely used extracting and separating agents. In this approach, an acidic or basic extracting agent is added to an aqueous solution that contains metal ions. Cueva et al. (2020) reported the extraction and enrichment of a vanadium- and tungsten-containing leach liquor obtained from spent SCR catalysts using Aliquat 336. This was followed by stripping at 55 °C with a mixture of NaOH and NaCl. Their research group enriched the solution seven times using a 0.5 mol L−1 Aliquat 336 solution. The resulting material was processed further via selective precipitation to obtain the title elements.2,3 Lozano et al. (2003) reported vanadium recovery and extraction characteristics obtained when extracting the metal from an aqueous sulphate solution using Primene R81 and Alamine 336.4 Ahn and Ahn (2008) studied vanadium extraction during solvent extraction of vanadium and titanium using Alamine 336. Although the extraction rate was approximately 40% at pH 1.0, it increased to more than 96% and >99.8% at pH ∼2.5 and 4.0, respectively.5 Gerhardt et al. (2001) reported extraction of 99.9% of tungsten using diisododecylamine (DIDA) and separation of tungsten, molybdenum, and rhenium from ore leachate.6 The separation and recovery of metal ions via solvent extraction method is limited in that one must select the extraction agent based on conditions such as the solution pH. Furthermore, separating the suspended materials and precipitates generated during extraction is difficult and requires extensive research.
Alternatively, separation and recovery of metal ions using an ion-exchange resin enables the reversible exchange of ions with the same charge between the ion bearing solution and the ion-exchange resin. Thus, separation can be performed based on the characteristics of the ions and the ion-exchange resin properties.7–11 This method has recently attracted significant attention, as it can selectively remove metal ions while reducing the solution contamination by impurities. In addition, since they can be regenerated after desorption, ion-exchange resins are economical to use. Furthermore, the adsorption and desorption preparation processes are simple.
After performing a study using ion-exchange resins, Hu et al. (2009) reported that 99.5% of vanadium was removed from a molybdate solution via adsorption using a strongly basic anion-exchange resin (D296).12 Ahn et al. (2003) reported the preparation of molybdate solution from a solution that contained molybdenum and tungsten.13 Adsorption experiments performed using chelating resins have demonstrated the possibility of effectively separating molybdenum and tungsten via adsorption in the 1.0 to 6.0 pH range.12,13
Some papers14,15 have described comprehensive biosorption investigations and highlighted that biosorption can bridge the gap between laboratory results and industrial applications. They have also demonstrated the importance of the green bio-sorption approach to applications such as toxic metal removal using the ion-exchange approach.16 Another study used phosphoryl-functionalised algal-PEI beads and algae biomass to adsorb rare earths and molybdenum.17,18 The agitation modes involved in mechanical, ultrasonic, and microwave techniques during uranium sorption using amine- and di-thizone-based magnetic chitosan hybrid materials were studied thoroughly.19 Using a 2-mercapto-benzimidazole derivative of chitosan as a sorbent reagent enabled the sonication effect to improve silver metal recovery.20 However, the recovery and separation of waste from vanadium and tungsten-containing selective catalytic reduction (SCR) denitrification catalysts via ion exchange has been rarely studied. In addition, detailed studies on the ion-exchange resin adsorption and desorption mechanisms of vanadium and tungsten and their related reaction formulas are insufficient. Therefore, in this study, we investigated the mechanisms by which vanadium and tungsten adsorb and desorb from mixed solutions of SCR denitrification waste catalysts and determined the factors that affect the reaction. The ion-exchange resin used in this study was a porous, strongly basic anion-exchange resin known to be effective in the adsorption of metal ions in solution.
Characteristics of the resin | Remarks |
---|---|
Trade name of the resin | MP 600 |
Supplier | Lanxess, Germany |
Ionic form as shipped | Cl− |
Functional group | Quaternary amine, type II |
Matrix | Cross linked polystyrene |
Structure | Macroporous |
Appearance | Beige, opaque |
Uniformity coefficient (max.) | 1.1 |
Mean bead size (mm) | 0.6 (±0.05) |
Bulk density (±5%) (g L−1) | 630 |
Density (approx. g mL−1) | 1.10 |
Water retention (wt%) | 5560 |
Total capacity (min. eq. L−1) | 1.1 |
Volume change (Cl → OH) (max. vol%) | 12 |
Stability (at pH-range) | 0 to 14 |
Storability of product (max. year) | 2 |
Storability temperature range (°C) | −20 to +40 |
BET surface area of resin | 22.125 m2 g−1 |
Single point surface area at P/Po | 21.272 m2 g−1 |
Langmuir surface area | 32.355 m2 g−1 |
Average pore volume | 0.1434 cm3 g−1 |
Average pore size | 25.928 nm |
In order to evaluate vanadium and tungsten ion adsorption and desorption characteristics using the ion-exchange resin, adsorption equilibrium experiments and adsorption rate experiments were carried out using the synthetic aqueous solution of ammonium metavanadate and ammonium tungstate pentahydrate in distilled water. Desorption experiments were performed based on the molar concentrations of HCl and NaOH.
The adsorption equilibrium experiment was carried out as a batch experiment in a shaking incubator. The quantity of ion-exchange resin was fixed and the adsorption rate was tested according to the initial acidity of the solution. An equimolar adsorption isotherm was measured with respect to the metal concentration. The metal adsorption rate was determined by treating 50 mL of the synthetic solution containing 500 mg L−1 each of vanadium and tungsten in the 1.0 to 13.0 pH range. This was followed by adding 0.3 g of ion-exchange resin and stirring at 303 K and 200 rpm for 1440 min. In the isothermal adsorption experiment, 0.5 g of resin was added to a solution containing vanadium and tungsten ions in the 50 to 500 mg L−1 concentration range. The mixture was stirred at 303 K and 200 rpm for 1440 min.
The adsorption rate tests were performed by adding 0.5 g of ion-exchange resin to 100 mL of a synthetic solution that contained 500 mg L−1 each of vanadium and tungsten, stirring for 1 to 1100 min, and collecting samples each hour.
Desorption experiments were performed to evaluate the desorption capabilities of the process as a function of the initial acidity and HCl or NaOH concentration. For the desorption experiment, 0.5 g of ion-exchange resin was added to a synthetic solution that contained 500 mg L−1 each of vanadium and tungsten and stirred for 1440 min. After reaching maximum adsorption, the ion-exchange resin was collected, washed with aqueous HCl (2 mol L−1) and NaOH (3 mol L−1) solution, and dried. For the process, the solution was stirred at 200 rpm and 303 K for 1 to 240 min and samples were taken at every hour.
The quantities of vanadium and tungsten ions adsorbed by the ion-exchange resin were determined by measuring the vanadium and tungsten concentrations after the adsorption reaction. We used eqn (1) to determine the volume of the solution and the amount of ion-exchange resin added.
![]() | (1) |
The resin XRD pattern (Fig. 2(B)) include a 2θ peak at 20° that indicates the mesoporous crystalline structure of the resin,21 which may help provide high adsorption of the target metal ions. Further, the XRD spectra taken after of adsorption at pH 1 and 7 exhibit significant changes in the resin XRD patterns with 2θ peaks at 10° and a broad peak at 20–30°. These may be related to adsorption of metal ions on the resin with a reduction22,23 or may involve a change in resin texture caused by conversion of amine groups.24 However, desorption of metals using 2 mol L−1 HCl and NaOH significantly influences the resin texture, as observed via the XRD spectrum. Desorption of metal using 2 mol L−1 NaOH seems to restore the original texture with small extra peaks at 2θ of 45° and 50° which helps reach the conclusion that there could be a multi-cycle reuse of the resin for the current process. The overall XRD results indicate that the resin is stable and retains its porous texture even after adsorption and desorption of both metal ions under different acidic conditions. Thus, we can restore the original resin structure using 2 mol L−1 NaOH (Fig. 2(B)) after desorption of metal ions either in acidic or basic conditions. Further, the FT-IR spectrum (Fig. 2(C)) confirms the resin structure.
In the FT-IR spectrum the broadband in the area of 3100–3500 cm−1 is ascribed to –N–H vibrations and the peak at 1490 cm−1 indicates –N+(CH2)3 moieties.25 Moreover, the sharp peak near 2900 cm−1 is ascribed the presence of the –C–C– stretching vibration in the resin. The peaks at 1200–1500 cm−1 are ascribed to the –C–N– bond frequency. The sharp peak at 1600 cm−1 indicates –CC– stretching vibrations. The peaks that appear in the area of 1000 to 800 cm−1 are ascribed to benzene ring –C
C– bands. After adsorption in acidic conditions or at pH 7.0, the positions of the peaks at 1200–1500 cm−1, 1600 cm−1, and in the area of 1000 to 800 cm−1 are altered. This indicates the loading of metal ions on the resin. These results confirm that metal ions are adsorbed predominantly via π–π bonding electrons rather than other functions on the surface of the resin. However, the functional groups are relocated to their original positions after desorption. The results are almost the same after desorption using 2 mol L−1 HCl and 2 mol L−1 NaOH. These results confirm the stability of the resin and that it can be reused multiple times.
The vanadium and tungsten adsorption behaviours vary with the initial acidity of the solution (Fig. 3(A)). Vanadium almost does not adsorb at pH 1.0 and 2.0. The adsorption rates of both ions decrease at pH ≥ 6.0, in addition to decreasing drastically in a strongly basic environment. Vanadium is present in the cationic form at pH 1.5, whereas it exists in the anionic form at pH > 1.5. Stable anionic vanadium adsorbs to anion-exchange resins quite easily in the 2.0–10.0 pH range. Tungsten shows a high adsorption rate in the acidic 1.0–6.0 pH range, but the adsorption rate tends to decrease at pH ≥ 9.0. The binding form of tungsten ions is obtained from the Eh-pH diagram as a function of solution acidity. The higher the basicity of the solution, the higher the ratio of WO42−. Various types of polyoxides appear if the acidity is increased by adding acid (pH decrease). Polyoxometalate anions appear as [W10O32]4− at pH 1.0–2.0, [H2W12O40]6− in the 2.0–4.0 pH range, and [W7O24]6− and [H2W12O42]10− in the 5.0–9.0 pH range. They precipitate as WO3·2H2O under strong acidic conditions (pH < 1.0). Therefore, tungsten exhibits optimum adsorption performance when it is present in the form of [H2W12O40]6− or [W7O24]6−.
Langmuir | Freundlich | Temkin | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
q (mg g−1) | b (L mg−1) | R2 | Pearson's correlation coefficient, r | KF ((mg g−1)(L mg−1)1/n) | 1/n | R2 | Pearson's correlation coefficient, r | ΔQ (J mol−1) | Ko (L mg−1) | R2 | Pearson's correlation coefficient, r | |
a Experimental conditions: initial pH is 7.0 using 0.5 g of ion exchange resin, temperature at 303 K and stirring is 200 rpm, reaction time is 1440 min, volume of the aqueous solution is 100 mL. | ||||||||||||
V | 75.301 | 1.564 | 0.974 | 0.993 | 53.469 | 0.078 | 0.644 | 0.845 | 3.337 × 102 | 3.786 × 107 | 0.770 | 0.877 |
W | 69.160 | 7.692 | 0.933 | 0.982 | 52.869 | 0.101 | 0.946 | 0.977 | 8.149 × 107 | 2.402 × 109 | 0.976 | 0.988 |
The Freundlich adsorption isotherm assumes that the adsorbent surface adsorbs to a multi-molecular layer considering the interaction between adsorbed metal ions with non-uniform surface energies. In this equation, KF, and n are constants that represent the suitability of the adsorption process. KF is the adsorption capacity, and 1/n is the slope. Respectively, they represent the adsorption strength and non-uniformity. The slope approaches zero as the non-uniformity increases.
The Temkin isotherm assumes a linear reduction in the heat of adsorption rather than a logarithmic drop, but disregards extremely high and low concentrations. It also implies that the bounding energy is distributed uniformly up to a certain maximum bonding energy. Here, Ko is the Temkin isotherm equilibrium constant (L g−1), T is the absolute temperature (K), R is the gas constant (8.314 J mol−1 K−1), and RT/ΔQ is the Temkin constant (J mol−1). The Langmuir, Freundlich, and Temkin adsorption isotherms are shown in Fig. S1† and the constants related to each adsorption isotherm are listed in Table 2. The suitability of the vanadium and tungsten data to isotherm model was evaluated using the correlation coefficient (R2) and statistical error, Pearson's correlation coefficient (r).
The Langmuir, Freundlich, and Temkin adsorption isotherm correlation coefficients (R2) for vanadium are 0.974, 0.644, and 0.770 respectively. This indicates that the Langmuir adsorption isotherm is the most suitable for explaining the vanadium ion isothermal adsorption experiment results. Further, the superior suitability of the Langmuir isotherm is confirmed via Pearson's correlation coefficient (r = 0.993), which is closer to 1 than those produced by the other isotherms in the case of vanadium. The Langmuir, Freundlich, and Temkin isothermal adsorption correlations and Pearson coefficients for tungsten all explain tungsten adsorption onto the resin well. However, the Freundlich and Temkin isotherms are confirmed to be most suitable based on their Pearson's correlation coefficients (r = 0.977 and 0.988 for the Freundlich and Temkin isotherms, respectively). The coefficients are close to 1 in the case of tungsten. Adsorption of vanadium ions to ion-exchange resins follows typical monolayer ion-exchange adsorption, whereas adsorption of tungsten ions is similar to a multimolecular adsorption by interacting ions by the poly-oxo-metalation of tungsten ions. According to the Eh-pH diagram, the proportion of WO42− in the solution increases as the basicity of the solution increases. The same concept applies to [W10O32]4− at pH 1 to 2, [H2W12O40]6− at pH 2 to 4, [W7O24]6− and [H2W12O42]10− at pH 5 to 9, and WO3·2H2O in strongly acidic solutions. Adsorption of the polyoxometallate ion affects the form in which tungsten adsorption occurs on the ion-exchange resin.
The suitability of the Freundlich adsorption isotherm is judged by the distribution of the 1/n value, as suggested by Fukukawa. Effective adsorption is possible when the 1/n value is in the range of 0.1 to 0.5. In this experiment, the 1/n value of vanadium is not suitable (0.078) and that of tungsten is 0.1007. This indicates that the adsorption of tungsten is explained well by the Freundlich adsorption isotherm.23,26–31
ln(qe − qt) = ln![]() | (2) |
Metal | Temperature, K | Pseudo 1st order model | Pseudo 2nd order model | Experimental, qe,exp, mg g−1 | |||||
---|---|---|---|---|---|---|---|---|---|
k1, min−1 | qe,cal, mg g−1 | R2 | k2, g mg−1 min−1 | qe,cal, mg g−1 | h, mg g−1 min−1 | R2 | |||
a Experimental conditions: initial pH is 7.0 using 0.1 g of ion exchange resin, temperature varied from 293 to 303 K, stirring is 200 rpm, volume of the aqueous solution is 100 mL, concentration of the metals in feed solution is 120 mg L−1 of vanadium (or) tungsten. | |||||||||
Vanadium | 293 | 5.97 × 10−3 | 33.418 | 0.922 | 8.06 × 10−4 | 63.412 | 3.244 | 0.999 | 62.460 |
303 | 3.74 × 10−3 | 22.365 | 0.454 | 1.31 × 10−3 | 67.340 | 5.953 | 0.999 | 68.060 | |
313 | 4.45 × 10−3 | 23.573 | 0.731 | 1.47 × 10−3 | 78.003 | 8.959 | 0.999 | 77.890 | |
Tungsten | 293 | 4.16 × 10−3 | 11.559 | 0.453 | 5.54 × 10−3 | 78.247 | 33.887 | 0.999 | 78.510 |
303 | 5.44 × 10−3 | 15.678 | 0.737 | 5.88 × 10−3 | 74.906 | 33.014 | 0.999 | 76.250 | |
313 | 1.57 × 10−3 | 7.162 | 0.102 | −7.28 × 10−3 | 69.396 | −35.075 | 0.999 | 73.470 |
A similar quadratic reaction rate model was proposed by Ho and Mckay based on the adsorption equilibrium capability31
![]() | (3) |
The k2, qe, and R2 values can be obtained from the slope and intercept of the above linear equation, where h (mg g−1 min−1) is the initial rate of adsorption.
h = k2qe2 | (4) |
The adsorption rate experiment results are shown in Fig. S2.† PSO kinetic parameter values are listed in Table 3 to enable understanding of the kinetic process.
The equilibrium adsorption qe,cal obtained from the similar PFO and PSO kinetic models and the equilibrium adsorption qe,exp obtained from the actual experiment are compared to the calculated values and qe,cal from the proposed models. The PSO model is found to be most suitable for representing the experimental results (Table 3). The R2 values indicate that the PSO model responds better to the experimental results as well.
ΔG = −RT![]() ![]() | (5) |
ΔG = ΔH − TΔS | (6) |
Upon combining eqn (5) and (6),
![]() | (7) |
The slope and intercept of the lnKe vs. 1000/T plot (Fig. S3†), which is based on the Van't Hoff relation (eqn (5)), correspond to the enthalpy and entropy, respectively. Table 4 shows the enthalpies, entropies, and Gibbs free energies associated with vanadium and tungsten adsorption. The enthalpy has a substantial positive value in the case of vanadium adsorption but a modest negative value in the case of tungsten adsorption. This indicates that the vanadium reaction is endothermic and leads to the realization that even a small temperature increase can increase the quantity of vanadium adsorbed on the resin. Tungsten adsorption, on the other hand, is somewhat exothermic. As shown in Fig. S3,† minor variations in the adsorption temperature have negligible influence on the adsorption process. The positive entropies noted in both cases show an increase in the randomness of the system in concordance to the second law of thermodynamics, which indicates that systems tend to increase their randomness spontaneously. These changes in entropy might occur due to structural changes happening between the metals and the resin during the adsorption process. Finally, the negative Gibbs free energy indicates thermodynamically favourable, spontaneous vanadium and tungsten adsorption processes.32–34
Thermodynamic parameter | Name of the metal | |
---|---|---|
Vanadium | Tungsten | |
a Experimental conditions: initial pH is 7.0 using 0.1 g of ion exchange resin, temperature varied from 293 to 303 K, stirring is 200 rpm, volume of the aqueous solution is 100 mL, concentration of the metals in feed solution is 120 mg L−1 of vanadium (or) tungsten. | ||
ΔH, kJ mol−1 | 19.39 | −1.78 |
ΔS, J K−1 mol−1 | 156.23 | 101.21 |
ΔG, kJ mol−1 at 298 K | −26.39 | −31.40 |
ΔG, kJ mol−1 at 303 K | −27.95 | −32.45 |
ΔG, kJ mol−1 at 313 K | −29.51 | −33.45 |
The developed adsorption and desorption methodology is compared to earlier reports on vanadium and tungsten adsorption in Table 5.35–43 The present metal removal and separation method is comparable to previously reported methods.
Name of the resin (or) adsorbent (or) sorbent | Name of the Metal (s) | Remarks | Reference |
---|---|---|---|
Amidoxime resin | Ga, V, Al | Vanadium was maximum adsorbed 26.32 mg g; at 50 °C temperature with 12 mol L−1 of NaOH solution was able to eluted 50% of vanadium | 35 |
Poly(styrene-divinylbenzene) (D860) resin and cross-linked acrylic acid (D314) resin | V, Al, Fe, P | Selective extraction of vanadium from 2000 mg L−1 of feed, eluted the 5505 mg L−1 after eight cycles of process with 30 mol L−1 of H2SO4 at 20 min time was reported | 36 |
Fabricated resin made with D314 (N(CH3)2 functional groups and cross-linked acrylate structure) and carbon nanotubes (CNTs) or activated carbon (AC) | V, Al, Fe, P, Si | D314 with activated carbon proved better ion selectivity | 37 |
Herein HZrO@D201, an adsorbent with decoration of nanosized hydrous zirconium oxide (HZrO) on anion exchange resin D201 | V, Cl−, NO3 −, SO42−, PO43− | V(V) was adsorbed maximum 118.1 mg g−1 at pH range 3 to 9. This study was performed for vanadium polluted ground water treatment | 38 |
Resin 201*7 | V, NH4+, Si, Al, Fe, Ca, Mg, K, Na | 99% of vanadium from 106 mg L−1 of feed solution was adsorbed with 1.6 g L−1 of resin at pH 6 to 8, temperature, 40 °C adsorption time requires 20 min (other associated elements were adsorbed less than 10%) | 39 |
Anion exchange resins (D201, D301, D314) and cation exchange resin (D860) were fabricated with activated carbon (AC) | V | D860/AC fabricated resin shown the highest adsorption capacity for V(V) at pH 2.0. Whereas other electrode i.e. D314/AC is the optimum due to worthy selective properties, compared with other electrodes | 40 |
The macro-porous resin D301 (analog to Amberlite IRA-9) | W, Mo | The resin D301 showed good selectivity in between title metals such as W and Mo | 41 and 42 |
The macro polyhydroxy chelating resin D403 (resembling Amberlite IRC-743) | W, Mo | W Adsorption was fits the Freundlich model, where as other metal Mo adsorption was fit Henry model. The reaction was endothermic | 43 |
Lewatit Monoplus MP600 | V, W | The developed methodology was applied to test real sample such as spent SCR catalyst. The results of applicability and res-usability studies concluded that the resin can be used for more than five cycles without losing its adsorption and re-useable efficacy more than 90% | Present method |
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1ra05253e |
This journal is © The Royal Society of Chemistry 2022 |