A Sustainable Methodology to Extract Bismuth from Secondary Sources

Abstract

Bismuth is a critical metal broadly used in cosmetic, medicine and metallurgy. However, its scarcity in Earth’s crust may impair further applications. An alternative is to recover bismuth from secondary sources. In this work, a methodology to extract bismuth from safety valves of discharged gas cylinders is proposed. Extraction and purification of bismuth were carried out using aqueous two-phase systems (ATPS) prepared with poly(ethylene oxide) polymer or poly(ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide) copolymer (L35) and electrolytes (NaNO₃, NH₄NO₃, sodium citrate or Na₂SO₄). The influence of iodide extractant concentration, tie-line length, electrolyte nature and polymer hydrophobicity on the bismuth extraction behavior were evaluated. Bismuth was significantly recovered in the absence of extractant, and its extraction depends on the ATPS composition, macromolecule and electrolyte nature. After three consecutive extractions bismuth was obtained with high purity (94.7%), resulting on a recovery of 38.5 g of bismuth per 1.00 kg of fusible plug, using L35 + NH₄NO₃ + H₂O ATPS, without any extractant.

Keywords:
bismuth; urban mining; aqueous two-phase system; critical metal; domestic waste; green separation

Introduction

Bismuth is considered a critical metal,¹ because it has a low occurrence in the Earth’s crust. However, it is mostly demanded by the cosmetic, pharmaceutical and metallurgical industries, mainly because of two reasons: low toxicity and low melting point.^2-6 The low toxicity of bismuth renders it more friendly to the environment and human beings, therefore enabling its ample use in the production of medicine, makeup and metallic alloys.^2,3,5,7-11 Its low melting point is also associated with its green nature, favoring the preparation of more sustainable metallic alloys that melt at relatively low temperatures.¹² Such alloys are employed in safety devices and systems for the detection and extinction of fires.^9,12

The high demand for bismuth has created a problem for the industries because of the limited distribution of the metal in the Earth’s crust. It is mainly obtained as a byproduct in the processing of other metallic ores, such as lead, copper, tungsten, tin and molybdenum.^13,14 Besides, China and Laos have dominated the export market of processed bismuth, since these countries produce 82 and 12%, respectively, of all bismuth that is consumed in the world.¹⁴ Because of this, countries like Brazil and the USA, among many others, depend highly on the importation of bismuth for the fabrication of different products, due to the small amount of bismuth that can be found in their soils.^14,15

One possible solution to overcome the lack of bismuth in several countries is to recover it from secondary sources, such as domestic and industrial residues, since the concentration of bismuth is much higher than that typically found in nature.¹⁶ However, the viability of this alternative process requires the development of techniques for the purification of bismuth from such secondary sources. It is a substantial scientific and technological challenge, because of the necessity to combine technical efficiency with economic feasibility and environmental safety.

Based on these questions, this study aimed to develop an efficient method to extract bismuth from safety valves of discharged gas cylinders using aqueous two-phase systems (ATPS). The high percentage of bismuth in these valves (as high as 50% in weight) and the massive consumption of bottled gas in the world, makes this material a potential candidate as a secondary source of bismuth.

Extraction and purification of bismuth were investigated in different ATPS composed by poly(ethylene oxide) polymer (PEO2000) or poly(ethylene oxide) polymer or poly(ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide) copolymer (L35) and electrolytes (NaNO₃, NH₄NO₃, C₆H₅Na₃O₇ or Na₂SO₄). The extraction behavior of bismuth was evaluated and optimized based on the following experimental parameters: ATPS-forming electrolyte, ATPS-forming macromolecule and tie-line length (TLL).

The ATPS proposed for the separation, and further purification assays can be applied in scale-up.^17,18 Moreover, the ATPS are both environmentally and economically sustainable, since their main component is water and the other components are nontoxic, cheap, biodegradable and recyclable.^19-21

Experimental

Materials

All chemicals used in this investigation were analytical grade. Deionized water produced by a Millipore Corp. deionizer (Massachusetts, USA) was used in the preparation of all solutions. The ATPS was prepared with macromolecule: triblock copolymer poly(ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide), denoted as L35, with an average molar mass of 1900 g mol^-1, or poly(ethylene oxide) with an average molar mass of 2000 g mol^-1, denoted as PEO2000. Both macromolecules and the bismuth salt BiN₃O₄·5H₂O were purchased from Aldrich (Wisconsin, USA). The chemicals NaNO₃, Na₂SO₄, KI, sodium citrate (C₆H₅Na₃O₇·2H₂O), CdCl₂·H₂O, SnCl₂·2H₂O, Pb(NO₃)₂ and Cu(NO₃)·3H₂O were supplied by Vetec (Rio de Janeiro, Brazil). HNO₃ and NH₄NO₃ were acquired from Merck (Darmstadt, Germany). The fusible plugs used as safety valves in gas cylinders were acquired in local Brazilian markets.

Equipment

A Metrohm pH-meter (model 827) was used to measure the pH of all solutions. Samples were weighed in a Shimadzu analytical balance (model AY 220), with an uncertainty of ± 0.0001 g. The ATPS was kept at 25.0 ± 0.1 °C in a Marconi water bath (model MA 184). In order to favor phase separation, a Thermo Scientific centrifuge (model Heraeus Megafuge 11R) was employed. The metal concentration was assessed with a flame atomic absorption spectrometer (FAAS; fabricated by Varian, model AA240) and/or a microwave-induced plasma atomic emission spectrometry (MP-AES; fabricated by Agilent Technologies Inc., model 4100). The operational conditions used in the FAAS and MP-AES assays are presented in Tables S1 and S2 (Supplementary Information (SI) section).

Preparation of the ATPS for extraction

The composition of each used ATPS (L35 + NaNO₃ + H₂O),²² (L35 + NH₄NO₃ + H₂O),²³ (L35 + C₆H₅Na₃O₇ + H₂O),²² (L35 + Na₂SO₄ + H₂O)²⁴ and (PEO2000 + NaNO₃ + H₂O)²⁵ were obtained from literature, together with their corresponding TLL.

An aqueous solution of 0.100 mol L^-1 HNO₃ (pH = 1.00) was used as a solvent for the preparation of stock solutions of macromolecule and electrolyte, whose concentrations are presented in Table 1. All ATPS was prepared to mix 2.00 g of the macromolecule stock solution with 2.00 g of the electrolyte stock solution.

Table 1
Concentrations of the stock solutions of macromolecules (M) and electrolytes (E) for the preparation of ATPS

ATPS	TLL / % (m/m)	M / % (m/m)	E / % (m/m)
L35 + NaNO3 + H2O	46.3	100a	29.3
L35 + NaNO3 + H2O	50.9	100a	30.8
L35 + NaNO3 + H2O	56.2	100a	32.4
L35 + NaNO3 + H2O	61.3	100a	34.1
L35 + NaNO3 + H2O	62.5	100a	36.2
L35 + NH4NO3 + H2O	50.4	100a	46.9
L35 + C6H5Na3O7 + H2O	48.2	100a	14.3
L35 + Na2SO4 + H2O	46.8	100a	11.0
PEO2000 + NaNO3 + H2O	41.2	63.2	51.0

• ^a
  Pure copolymer. ATPS: aqueous two-phase systems; TLL: tie-line length; M: macromolecule; E: electrolyte; L35: poly(ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide) copolymer; PEO2000: poly(ethylene oxide) polymer.

Effect of the amount of extractant agent on the recovery of bismuth

The L35 + NaNO₃ + H₂O ATPS, TLL = 46.3% (m/m) was used to investigate the influence of the quantity of extractant added on the recovery of bismuth ions. Work solutions of an iodide (I^-) extractant agent were prepared by adding different amounts of the potassium iodide (0.86, 4.30, 8.60, 17.2 and 34.5 μmol) to L35 [100% (m/m)]. Work solutions of Bi^III (concentration ranging between 9.00 and 108 mg kg^-1) were prepared by dissolving Bi^III in the NaNO₃ stock solution [29.3% (m/m)]. In a centrifuge tube, 2.00 g of Bi^III work solution were mixed with 2.00 g of the iodide work solution. The tube was manually stirred for 3 min, centrifuged at 10000 × g for 15 min and then kept in a water bath at 25.0 °C for 10 min. After these steps, the system separated in two clear phases, named as the upper phase (UP) and the bottom phase (BP). Aliquots of UP and BP were collected and diluted 10 times to determine the concentration of bismuth by FAAS (Table S1, SI section). The extraction percentage of bismuth (E_Bi) in all experiments was obtained by equation 1:

(1) $E_{\mathrm{Bi}}\left(\%\right)=\frac{n_{\mathrm{Bi}}^{\mathrm{UP}}}{n_{\mathrm{Bi}}^T}\times 100$

where is the molar amount of bismuth in the upper phase and is the total molar amount of bismuth in the system. All experiments were carried out in triplicate.

Effect of the ATPS-forming electrolyte on the recovery of bismuth

The ATPS formed by L35 + NaNO₃ + H₂O [TLL = 46.3% (m/m)], L35 + NaNO₃ + H₂O [TLL = 50.9% (m/m)], L35 + NH₄NO₃ + H₂O [TLL = 50.4% (m/m)], L35 + C₆H₅Na₃O₇ + H₂O [TLL = 48.2% (m/m)] and L35 + Na₂SO₄ + H₂O [TLL = 46.8% (m/m)] were used to study the effect of the ATPS-forming electrolyte on the recovery of bismuth.

Bi^III work solutions with concentrations of 9.00, 18.0, 27.0, 36.0, 72.0, 90.0 and 108 mg kg^-1 were prepared by dissolving the bismuth salt in the following electrolyte stock solutions: NaNO₃ [29.3% (m/m)], NaNO₃ [30.8% (m/m)], Na₂SO₄ [11.0% (m/m)], NH₄NO₃ [46.9% (m/m)] or C₆H₅Na₃O₇ [14.3% (m/m)]. In a centrifuge tube, 2.00 g of the bismuth work solution were mixed with 2.00 g of L35. Again, the tube was manually stirred for 3 min, centrifuged at 10000 × g for 15 min and then kept in a water bath at 25.0 °C for 10 min, after which two clear phases were formed. Aliquots of UP and BP were collected and diluted 10 times to determine the concentration of bismuth by FAAS (Table S1, SI section), then calculating E_Bi with equation 1. All experiments were carried out in triplicate.

Effect of the hydrophobicity of the ATPS-forming macromolecule on the recovery of bismuth

The extraction behavior of bismuth was evaluated in systems with distinct macromolecules [L35 + NaNO₃ + H₂O, TLL = 46.3% (m/m) and PEO2000 + NaNO₃ + H₂O, TLL = 41.2% (m/m)]. Bi^III work solutions with concentrations of 9.00, 18.0, 27.0, 36.0, 72.0, 90.0 and 108 mg kg^-1 were prepared by dissolving the bismuth salt in the stock solutions of the NaNO₃ electrolyte at concentrations of 29.3 or 51.0% (m/m). In centrifuge tubes, two different systems were prepared. First, 2.00 g of the PEO2000 63.2% (m/m) stock solution were mixed with 2.00 g of the Bi^III work solution in NaNO₃ 51.0% (m/m). Then, 2.00 g of the L35 stock solution were mixed with 2.00 g of the Bi^III work solution in NaNO₃ 29.3% (m/m). Again, the tube was manually stirred for 3 min, centrifuged at 10000 × g for 15 min and then kept in a water bath at 25.0 °C for 10 min, after which two clear phases were formed. Aliquots of UP and BP were collected and diluted 10 times to determine the concentration of bismuth by FAAS (Table S1, SI section), then calculating E_Bi with equation 1. All experiments were carried out in triplicate.

Effect of the ATPS composition on the recovery of bismuth

The L35 + NaNO₃ + H₂O ATPS in different TLL [46.3, 50.9, 56.2, 61.3 and 62.5% (m/m)] was used to study the effect of the ATPS composition on the recovery of bismuth.

Bi^III work solutions with concentrations of 9.00, 18.0, 27.0, 36.0, 72.0, 90.0 and 108 mg kg^-1 were prepared by dissolving the bismuth salt in the stock solutions of the NaNO₃ electrolyte at different concentrations: 29.3, 30.8, 32.4, 34.1 and 36.2% (m/m). In a centrifuge tube, 2.00 g of the bismuth work solution were mixed with 2.00 g of L35. Again, the tube was manually stirred for 3 min, centrifuged at 10000 × g for 15 min and then kept in a water bath at 25.0 °C for 10 min, after which two clear phases were formed. Aliquots of UP and BP were collected and diluted 10 times to determine the concentration of bismuth by FAAS (Table S1, SI section), then calculating E_Bi with equation 1. All experiments were carried out in triplicate.

Leaching of the fusible plug sample

The fusible plug sample was previously cleaned and dried, and the initial mass (12.0 g) was measured in the analytical balance. The plug was added with 39.7 mL of a 7.00 mol L^-1 HNO₃ solution to a two-neck round-bottom flask. The mixture was heated at 70 °C for 80 min with a heating mantle. After cooling, the mixture was filtered out, and the leachate was collected as a clear solution. The leachate was diluted 2000 times to determine the amount of bismuth, copper, lead, cadmium and tin using an MP-AES (Table S2, SI section) with the help of a multi-element analytical curve [coefficient of determination (R²) ≥ 0.999]. All experiments were carried out in triplicate. This leachate without dilution was stored for future bismuth purification using ATPS.

Recovery and purification of bismuth from the fusible plug

The recovery of bismuth was carried out by three consecutive extraction steps using ATPS. For the first extraction step, 4.21 g of the leachate, 4.20 g of L35, 4.93 g of NH₄NO₃ and 1.75 g of a 0.100 mol L^-1 HNO₃ solution were mixed in a centrifuge tube. Then, the tube was manually stirred for 3 min, centrifuged at 3000 × g for 15 min and kept in a water bath at 25.0 °C for 10 min. After that, it was obtained two clear phases. This system was then named ATPS1. For the second extraction step, 5.00 g of the UP of the ATPS1 were transferred to a new centrifuge tube. In this tube, 0.0330 g of L35, 2.47 g of NH₄NO₃, and 2.49 g of a 0.100 mol L^-1 HNO₃ solution were added. The tube was manually stirred for 3 min, centrifuged at 3000 × g for 15 min and kept in a water bath at 25.0 °C for 10 min. So, it was obtained two clear phases. This system was named ATPS2. Finally, in the third extraction step, it was transferred 2.00 g of the UP of ATPS2, 0.013 g of L35, 0.970 g of NH₄NO₃ and 1.00 g of a 0.100 mol L^-1 HNO₃ solution to another centrifuge tube. This final system was manually stirred for 3 min, centrifuged at 3000 × g for 15 min and kept in a water bath at 25.0 °C for 10 min. Then, it was obtained two clear phases. This system was then named ATPS3. Aliquots of the UP and BP of ATPS1, ATPS2 and ATPS3 were collected and adequately diluted to determine bismuth, copper, lead, cadmium and tin by FAAS or MP-EAS (Tables S1 and S2, SI section). All experiments were carried out in triplicate.

Results and Discussion

Effect of the amount of extractant agent on the recovery of bismuth

It has been demonstrated that ATPS can be very effective in the extraction of metallic ions.^26-30 The extraction procedures are usually conducted with the aid of an extractant agent, whose main function is to enhance the extraction of ions to the polymer-rich phase.³¹ Because of this, it is important to determine optimal conditions for the use of such agent, with the choice of the most appropriate chemical and the selection of the best concentration, which can favor the extraction or separation of metals in a given sample.^29-32

Bulgariu and Bulgariu³³ investigated the extraction behavior of bismuth in the ATPS formed by PEG1500 + (NH₄)₂SO₄ + H₂O, in the presence of four inorganic extractants (I^-, Br^-, Cl^- and SCN^-). The efficiency of extractants follows the order: I^- > Br^- > Cl^- > SCN^-. Based on these results, in the present work, the extraction behavior of bismuth was investigated in a different system consisting of L35 + NaNO₃ + H₂O in the presence of the most promising extractant, I^-.

The values of E_Bi obtained when adding different amounts of iodide (I^-) extractant agent to the L35 + NaNO₃ + H₂O ATPS in the TLL = 46.3% (m/m), at 25.0 °C and pH 1.00 are presented in Figure 1. In the absence of I^-, E_Bi was 79.1 ± 5.9%. When adding I^- at concentrations higher than 2.15 μmol, the value of E_Bi increased to approximately 98%.

Figure 1
Effect of the amount of the iodide (I^-) extractant agent on the extraction percentage of bismuth (E_Bi) using the L35 + NaNO₃ + H₂O ATPS, with TLL = 46.3% (m/m), at 25.0 °C and pH = 1.00.

In aqueous solution, bismuth ions can undergo a series of hydrolysis reactions and polynucleation processes which justify the performance of the extraction procedures in the absence of extractant agents. In this case, E_Bi was as high as 79.1 ± 5.9%. Investigations show that the species [Bi₆O₄(OH)₄]⁶⁺ and [Bi₆O₅(OH)₃]⁵⁺ are favored in acidic media. Equilibrium between these species (equation 2) depends on the pH of the medium, which at low pH the compound [Bi₆O₄(OH)₄](NO₃)₆.yH₂O (y = 1 or 4) is formed, and at slightly higher pH the compound [Bi₆O₅(OH)₃](NO₃)₅.3H₂O is formed.^34,35

(2) ${\left[{\mathrm{Bi}}_6{O}_4{\left(\mathrm{OH}\right)}_4\right]}_{\left(\mathrm{aq}\right)}^{6+}+H_2{O}_{\left(l\right)}\rightleftharpoons {\left[{\mathrm{Bi}}_6{O}_5{\left(\mathrm{OH}\right)}_3\right]}_{\left(\mathrm{aq}\right)}^{5+}+H_3{O}_{\left(\mathrm{aq}\right)}^{+}$

Based on calorimetric measurements, da Silva and Loh³⁶ proposed that the UP of ATPS contains macromolecules and cations of the system-forming electrolyte adsorbed on the polymeric chain, generating a positively-charged surface (pseudopolycation) that interacts with negatively-charged species. Therefore, pseudopolycations found in the UP of L35 + NaNO₃ + H₂O ATPS must establish repulsive interactions with the bismuth cationic species that are also present in that phase. However, specific interactions among the bismuth polymorphous species and the macromolecule chain must overcome such repulsive phenomena.

The addition of small amounts of iodide to the ATPS increased the value of E_Bi due to the formation of anionic complexes between I^- and Bi^III, as indicated by equation 3:

(3) ${\mathrm{Bi}}_{\left(\mathrm{aq}\right)}^{3+}+{\mathrm{nI}}_{\left(\mathrm{aq}\right)}^{-}\rightleftharpoons {\left[{\mathrm{BiI}}_n\right]}_{\left(\mathrm{aq}\right)}^{3-n}$

High values of formation constants revealed that the complexes are stable (log K₁ = 3.63, log K₄ = 14.95, log K₅ = 16.80 and log K₆ = 18.80).³⁷ The transfer of bismuth species to the UP in the presence of iodide occurs because of electrostatic interactions between the bismuth anionic complexes and the pseudopolycations that are abundant in that phase. As these complexes are formed, they are transferred to the UP from the BP, a phenomenon which induces the formation of more complexes in the BP through equilibrium shifts. When adding iodide, the equilibrium is also shifted until reaching a saturation point, after which the addition of more iodide has no further effect on E_Bi. From 2.15 μmol of added iodide, the value of E_Bi was stabilized close to 98%. The relatively small increment in E_Bi in the presence of iodide, from ca. 79 to ca. 98%, demonstrates that it is not necessary to employ an extractant agent to obtain significant recovery of bismuth with ATPS. Besides being environmentally advantageous, the elimination of such an extractant agent reduces costs in the recovery of bismuth. For these reasons, further investigations in this work have been conducted without any extractant agent in the systems.

Effect of the ATPS-forming cation on the recovery of bismuth

The extraction percentage of metallic ions in ATPS normally depends on the nature of the cation of the ATPS-forming electrolyte.^26,29,38,39 The contribution of the cation to the distribution of bismuth between phases of the ATPS can be viewed in Figure 2. Figure 2 shows the E_Bi as a function of bismuth concentration for the systems L35 + NaNO₃ + H₂O (TLL = 50.9% (m/m)) and L35 + NH₄NO₃ + H₂O (TLL = 50.4% (m/m)) at 25.0 °C, pH = 1.00, and in the absence of extractant agent.

Figure 2
Effect of ATPS-forming cation on the extraction percentage of bismuth (E_Bi). ATPS: (□) L35 + NaNO₃ + H₂O (TLL = 50.9% (m/m)) and (○) L35 + NH₄NO₃ + H₂O (TLL = 50.4% (m/m)), at 25.0 °C, pH = 1.00, in the absence of extractant agent.

The ATPS formed with NaNO₃ was more efficient in the partition of bismuth from the BP to the UP than the ATPS formed with NH₄NO₃ within all the examined bismuth concentration range. ATPS formed with sodium cations promoted a maximal E_Bi of 83.2 ± 1.9%, while ATPS formed with ammonium cations has a maximal E_Bi of 77.8 ± 0.3%. Since these electrolytes have the same nitrate anion, one can describe the variation in E_Bi to the cation. Calorimetric investigations reported elsewhere³⁶ suggest that pseudopolycations formed with ammonium would be less positively charged than those formed with sodium cations. Therefore, the ATPS formed with NaNO₃ would be more efficient in repelling bismuth species, which would cause a decrease in their E_Bi when compared to those of ATPS formed with NH₄NO₃. However, the concentration of sodium cations in the UP of the L35 + NaNO₃ + H₂O ATPS is 13.19% (m/m), whilst the concentration of ammonium cations in the L35 + NH₄NO₃ + H₂O ATPS is 23.98% (m/m).^22,23 The higher concentration of ammonium cations in the UP of the L35 + NH₄NO₃ + H₂O ATPS enhances repulsive interactions between the bismuth polymorphous cationic species and the pseudopolycations in this ATPS. As a consequence, the extraction of bismuth in the ATPS containing ammonium is a little lower than that promoted by the sodium-based ATPS.

Effect of the ATPS-forming anion on the recovery of bismuth

The contribution of the anion to the extraction of bismuth was also examined. Figure 3 shows the values of E_Bi as a function of bismuth concentration for ATPS with L35 and sodium salt with different anions (nitrate, sulfate and citrate) in the TLL ca. 47% (m/m), pH = 1.00, 25.0 °C, and absence of extractant agent.

Figure 3
Effect of ATPS-forming anion on the extraction percentage of bismuth (E_Bi). ATPS: (■) L35 + NaNO₃ + H₂O (TLL = 46.3% (m/m)); (●) L35 + Na₂SO₄ + H₂O (TLL = 46.7% (m/m)) and (▽) L35 + C₆H₅Na₃O₇ + H₂O (TLL = 48.1% (m/m)), at 25.0 °C, pH = 1.00, in the absence of extractant agent.

As opposed to the effect of cation, the nature of the anion profoundly impacted the values of E_Bi. The maximal E_Bi for the ATPS formed with nitrate was 78.8 ± 0.5%, while with sulfate and citrate the maximal E_Bi was only 1.49 ± 0.59 and 2.43 ± 1.30%, respectively. ATPS formed with sulfate and citrate anions were less efficient in the extraction of bismuth because the formation of stable bismuth complexes with sulfate⁴⁰ and citrate⁴¹ is more favorable in the BP of the ATPS that is an electrolyte-rich phase.

In order to better explore the contribution of the anion to E_Bi, known amounts of nitrate were added to the L35 + Na₂SO₄ + H₂O ATPS (Figure 4a) and known amounts of sulfate were added to the L35 + NaNO₃ + H₂O ATPS (Figure 4b).

The addition of nitrate ions to the L35 + Na₂SO₄ + H₂O ATPS could not enhance the extraction of bismuth (E_Bi < 2.00%). On the other hand, the continuous increment of sulfate ions to the system L35 + NaNO₃ + H₂O reduced E_Bi from 80.0 ± 0.5% in the absence of sulfate to 67.7 ± 0.3% when the [SO₄^2-:Bi] molar ratio was 500:1. These data confirm the stability of bismuth sulfate complexes (log K₁ = 1.98, log K₂ = 3.41, log K₃ = 4.08, log K₄ = 4.34, log K₅ = 4.6)⁴¹ and their preferential interaction with the components of the electrolyte-rich phase (BP). The ATPS formed with NaNO₃ in the absence of sulfate proved to be a more efficient system to extract bismuth.

Figure 4
Extraction of bismuth in the presence of nitrate and sulfate ions. ATPS: (a) (●) L35 + Na₂SO₄ + H₂O (TLL = 46.7% (m/m)) in the presence of nitrate; (b) (■) L35 + NaNO₃ + H₂O (TLL = 46.3% (m/m)) in the presence of sulfate, at 25.0 °C and pH = 1.00.

Effect of the hydrophobicity of the ATPS-forming macromolecule on the recovery of bismuth

In this particular investigation, two macromolecules, L35 and PE02000, with different hydrophobicity were investigated. The NaNO₃ was selected as the ATPS-forming electrolyte because ATPS containing this sodium salt proved to be more useful to extract bismuth in previous studies. The ATPS formed with L35 is more hydrophobic because of the polypropylene oxide (PPO) and polyethylene oxide (PEO) segments of the macromolecule than PEO2000, which is more hydrophilic because it contains only PEO segments.

Figure 5 shows the values of E_Bi as a function of bismuth concentration for the PEO2000 + NaNO₃ + H₂O ATPS, TLL = 41.2% (m/m), and L35 + NaNO₃ + H₂O, TLL = 46.3% (m/m), at pH = 1.00, 25 °C, and absence of extractant agent.

Figure 5
Effect of the hydrophobicity of the ATPS-forming macromolecule on the extraction percentage of bismuth (E_Bi). ATPS: (■) L35 + NaNO₃ + H₂O (TLL = 46.3% (m/m)); (★) PEO2000 + NaNO₃ + H₂O (TLL = 41.2% (m/m)), at 25 °C and pH = 1.00, in the absence of extractant agent.

The maximal value of E_Bi obtained with the L35 ATPS was 85.0 ± 0.5%, while for the PEO ATPS was 98.3 ± 2.4%. The values of E_Bi are indicative that bismuth polymorphous cationic complexes preferably interact with the ethylene oxide (EO) segments of the macromolecules. Therefore, the lower amount of EO segments on the L35 molecule (50 wt.%) when compared to PEO2000 (100 wt.%) justifies the lower E_Bi of L35 ATPS. ATPS with L35 was selected to carry out further investigations and to develop the methodology to recover bismuth from the discharged safety valves of gas cylinders. As shown in Figure 6, when ATPS with L35 + NaNO₃ was used for bismuth recuperation from the discharged safety valve the E_Bi was quantitative. Also, the choice of L35 is relevant to scale-up the extraction and separation process with no significant negative impact on its overall efficiency. Triblock copolymers are thermosensitive, presenting a lower cloud point temperature than PEO polymer.²³ Consequently, L35 macromolecules are easier to be recycled by the phase separation induced by the temperature increase.²³

Figure 6
Extraction of metals found in the leachates of the fusible plugs of gas cylinders using ATPS. (a) L35 + NaNO₃ + H₂O, TLL = 62.6% (m/m) and (b) L35 + NH₄NO₃ + H₂O, TLL = 50.4% (m/m), at 25 °C, in the absence of extractant agent.

Influence of the ATPS composition on the recovery of bismuth

Their intensive thermodynamical properties govern the distribution of a given solute between both ATPS phases. The TLL, associated with the ATPS composition, numerically expresses the difference between these properties, and can be calculated with equation 4:

(4) $\mathrm{TLL}={\left[{\left(C_M^{\mathrm{UP}}-C_M^{\mathrm{BP}}\right)}^2+{\left(C_E^{\mathrm{UP}}-C_E^{\mathrm{BP}}\right)}^2\right]}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}$

where and are the concentrations of macromolecule in the upper and bottom phase, respectively, and and are the concentrations of electrolyte in the upper and bottom phase, respectively. All concentrations in this work are expressed in % (m/m).

The bismuth extraction percentage was determined for different TLL with the L35 + NaNO₃ + H₂O ATPS (Figure 7).

Figure 7
Effect of the tie-line length of the L35 + NaNO₃ + H₂O ATPS on the extraction percentage of bismuth (E_Bi). (■) TLL = 46.3% (m/m); (□) CLA = 50.9% (m/m); (⬪) TLL = 56.2% (m/m); (⬨) TLL = 61.3% (m/m); (☆) TLL = 62.6% (m/m), at 25 °C and pH = 1.00, in the absence of extractant agent.

The average E_Bi values obtained with TLL 46.3, 50.9, 56.2, 61.3 and 62.6% (m/m) were 76.5 ± 1.6, 81.5 ± 1.4, 88.4 ± 2.0, 86.8 ± 2.9 and 86.3 ± 3.6%, respectively. A slight increase in E_Bi is influenced with increasing TLL, reaching a plateau from TLL = 56.2% (m/m), regardless of the amount of bismuth added to the ATPS. Such extraction behavior is attributed to the increase in the difference between the intensive thermodynamically properties as TLL increases. The small variation in the concentration of NaNO₃ in the UP of the ATPS²² helps to justify the fact that E_Bi remains constant from TLL = 56.2% (m/m), with the formation of pseudopolycations that have similar charge densities.

Recovery and purification of bismuth from safety valves of gas cylinders (fusible plugs)

ATPS has been employed in the recovery and purification of bismuth from waste fusible plugs of gas cylinders. Fusible plugs are metallic alloys (Wood’s alloy) composed mainly of bismuth, cadmium, lead and tin, covered with a copper casing.⁴² The plug + copper casing assembly underwent leaching treatment with nitric acid, whereby only the fusible plug was solubilized, and thereby separated from the copper casing. The concentration of metals in the fusible plug was determined as: Bi = 14.8 ± 0.6 g kg^-1; Pb = 10.8 ± 0.5 g kg^-1; Cu = 51.4 ± 0.9 g kg^-1; Cd = 1.60 ± 0.01 g kg^-1; and Sn = 0.255 ± 0.003 g kg^-1.

Figure 6 presents the extraction percentages (E) of these metals using the L35 + NaNO₃ + H₂O ATPS, TLL = 62.6% (m/m), and L35 + NH₄NO₃ + H₂O, TLL = 50.4% (m/m). The L35 + NaNO₃ + H₂O ATPS was more efficient, as already discussed earlier.

The presence of concomitant metals does not affect the E_Bi that is 106 ± 4%, but high extraction percentages are obtained for the concomitant metals: E_Cu = 77.1 ± 4.0%; E_Pb = 42.5 ± 0.8%; E_Cd = 21.1 ± 2.3%; and E_Sn = 14.0 ± 0.7%. It compromises the purity of the recovered bismuth samples, which was only 26.1%. Similar behavior is observed when the L35 + NH₄NO₃ + H₂O ATPS is applied in the extraction of bismuth. In this case, E_Bi was 91.7 ± 5.7%, but the extraction percentages of Cu, Pb, Cd and Sn were lower: 23.7 ± 0.0, 37.9 ± 0.6, 16.6 ± 0.5 and 3.96 ± 0.67%, respectively. Therefore, bismuth could be recovered with a higher purity of 45.1% with this ATPS.

However, 45.1% purity is not an appropriate level for a given product to be available to the industrial production chain. As a consequence, three consecutive extraction procedures were carried out to obtain bismuth with higher purity. The extraction percentages (E) of all metals found in the fusible plug leachate after three consecutive steps are depicted in Figure 8 for the L35 + NH₄NO₃ + H₂O ATPS.

Figure 8
Extraction of metals found in the leachates of the fusible plugs of gas cylinders after three consecutive extraction steps using the L35 + NH₄NO₃ + H₂O ATPS, TLL = 50.4% (m/m) at 25 °C, in the absence of extractant agent.

It can be observed that, although E_Bi was reduced to 61.9 ± 2.1%, the bismuth was obtained with very high purity (94.7%), resulting from the recovery of 38.5 g of bismuth per 1.00 kg of the fusible plug.

Conclusions

In this work, it has been demonstrated that aqueous two-phase systems are useful for the extraction of bismuth from secondary sources without any extractant agent or organic solvents. The partition of bismuth to the macromolecule-rich phase (upper phase of ATPS) was governed by specific interactions between the bismuth species and the ethylene oxide group of the macromolecules. The nature of the ATPS-forming electrolyte, the nature of the macromolecule and the composition of the ATPS (expressed by the corresponding TLL) significantly affect the recovery of bismuth. Although some investigated ATPS are useful to extract bismuth, the ATPS formed with L35 + NH₄NO₃ + H₂O and TLL = 50.4% (m/m), working at 25.0 °C in the absence of extractant agent, was the one that promoted the highest recovery of bismuth with high-purity (94.7%). This work is a pioneer in opening possibilities to implement ATPS in the extraction of such a critical metal as bismuth on a large scale. Therefore, it is possible to carry out urban mining operations to enhance the production of bismuth, with distinct environmental and economic advantages to society. The following environmental and economic advances have been reached: (i) the proposed methodology to obtain bismuth is green and economical, since the ATPS is composed mainly of water, also comprising other nontoxic, cheap, biodegradable and recyclable chemicals that are found in relatively low concentrations; (ii) the discharge of fusible plugs of gas cylinders could be diminished and even avoided, and at the same time some aggregate value would be granted to such discharged material; (iii) natural reserves could be preserved, because bismuth would not be acquired or extracted from the nature itself; (iv) countries with little or no potential to produce bismuth because of scarce natural sources would be more independent on the production and use of this critical metal.

Supplementary Information

Supplementary data are available free of charge at http://jbcs.sbq.org.br as PDF file.

Acknowledgments

The authors are indebted to the Instituto Nacional de Ciências e Tecnologias Analíticas Avançadas (INCTAA), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG), and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) for granting financial support to this research. S. J. R. V. and T. C. S. R. also acknowledge CAPES for granting studentships. M. C. H. and L. H. M. S. also acknowledge CNPq for granting scholarships.

References

• ¹
  Panousi, S.; Harper, E. M.; Nuss, P.; Eckelman, M. J.; Hakimian, A.; Graedel, T. E.; J. Ind. Ecol. 2016, 20, 837.
• ²
  Liu, X.; Xiao, M.; Xu, L.; Miao, Y.; Ouyang, R.; J. Nanosci. Nanotechnol. 2016, 16, 6679.
• ³
  Zaccariello, G.; Back, M.; Zanello, M.; Canton, P.; ACS Appl. Mater. Interfaces 2017, 9, 1913.
• ⁴
  Abdel-Fadeel, M. A.; Al-Saidi, H. M.; El-Bindary, A. A.; El-Sonbati, A. Z.; Alharthi, S. S.; J. Mol. Liq. 2018, 249, 963.
• ⁵
  Al-Saidi, H. M.; Abdel-Fadeel, M. A.; El-Sonbati, A. Z.; El-Bindary, A. A.; J. Mol. Liq. 2016, 216, 693.
• ⁶
  Amin, A. S.; Zaafarany, I. A.; J. Taibah Univ. Sci. 2015, 9, 490.
• ⁷
  Filella, M.; J. Environ. Monit. 2010, 12, 90.
• ⁸
  Afkhami, A.; Madrakian, T.; Siampour, H.; J. Braz. Chem. Soc. 2006, 17, 797.
• ⁹
  Padmanabha, S. K.; Rao, S. V.; Rasayan J. Chem. 2011, 4, 857.
• ¹⁰
  Wang, R.; Lai, T. P.; Gao, P.; Zhang, H.; Ho, P.-L.; Woo, P. C.-Y.; Ma, G.; Kao, R. Y.-T.; Li, H.; Sun, H.; Nat. Commun. 2018, 9, 439.
• ¹¹
  Campos, V.; Almaguer-Flores, A.; Velasco-Aria, D.; Díaz, D.; Rodil, S.; Mater. Sci. Eng., A 2018, 8, 142.
• ¹²
  Kamal, M.; Mazen, S.; El-Bediwi, A. B.; Kashita, E.; Radiat. Eff. Defects Solids 2005, 160, 369.
• ¹³
  Chagnon, M. J. In Kirk-Othmer Encyclopedia of Chemical Technology, 5^th ed.; Kirk-Othmer; Kroschwitz, J. I.; Seidel, A., eds; Wiley: Hoboken, New Jersey, USA, 2007, p. 1.
• ¹⁴
  https://minerals.usgs.gov/minerals/pubs/commodity/bismuth/myb1-2016-bismu.pdf, accessed on December 20, 2018.
  » https://minerals.usgs.gov/minerals/pubs/commodity/bismuth/myb1-2016-bismu.pdf
• ¹⁵
  Guerra, W.; Alves, F. E.; Silva, K. C. C.; Quim. Nova Esc. 2011, 33, 193.
• ¹⁶
  Campos, K.; Domingo, R.; Vincent, T.; Ruiz, M.; Sastre, A. M.; Guibal, E.; Water Res. 2008, 42, 4019.
• ¹⁷
  Torres-Acosta, M. A.; Mayolo-Deloisa, K.; Gonzalez-Valdez, J.; Rito-Palomares, M.; Biotechnol. J. 2019, 14, e1800117.
• ¹⁸
  Sutherland, I.; Hewitson, P.; Siebers, R.; van den Heuvel, R.; Arbenz, L.; Kinkel, J.; Fisher, D.; J. Chromatogr. A 2011, 1218, 5527.
• ¹⁹
  Xie, X.; Yan, Y.; Han, J.; Wang, Y.; Yin, G.; Guan, W.; J. Chem. Eng. Data 2010, 55, 2857.
• ²⁰
  Persson, J.; Kaul, A.; Tjerneld, F.; J. Chromatogr. B 2000, 743, 115.
• ²¹
  Persson, J.; Johansson, H.-O.; Tjerneld, F.; J. Chromatogr. A 1999, 864, 31.
• ²²
  Martins, J. P.; Mageste, A. B.; da Silva, M. D. H.; da Silva, L. H. M.; Patrício, P. D.; Coimbra, J. S. D.; Minim, L. A.; J. Chem. Eng. Data 2009, 54, 2891.
• ²³
  Veloso, A. C. G.; Patricio, P. R.; Quintao, J. C.; de Carvalho, R. M. M.; da Silva, L. H. M.; Hespanhol, M. C.; Fluid Phase Equilib. 2018, 469, 26.
• ²⁴
  da Silva, M. D. H.; da Silva, L. H. M.; Amim, J.; Guimarães, R. O.; Martins, J. P.; J. Chem. Eng. Data 2006, 51, 2260.
• ²⁵
  Graber, T. A.; Taboada, M. E.; Asenjo, J. K.; Andrews, B. A.; J. Chem. Eng. Data 2001, 46, 765.
• ²⁶
  Patricio, P. D.; Mesquita, M. C.; da Silva, L. H. M.; da Silva, M. C. H.; J. Hazard. Mater. 2011, 193, 311.
• ²⁷
  Rodrigues, G. D.; de Lemos, L. R.; da Silva, L. H.; da Silva, M. C. H.; J. Chromatogr. A 2013, 1279, 13.
• ²⁸
  Rodrigues, G. D.; da Silva, M. D. H.; da Silva, L. H. M.; Paggioli, F. J.; Minim, L. A.; Coimbra, J. S. D.; Sep. Purif. Technol. 2008, 62, 687.
• ²⁹
  de Lemos, L. R.; Campos, R. A.; Rodrigues, G. D.; da Silva, L. H. M.; da Silva, M. C. H.; Sep. Purif. Technol. 2013, 115, 107.
• ³⁰
  da Cunha, R. C.; Patricio, P. R.; Vargas, S. J. R.; da Silva, L. H. M.; da Silva, M. C. H.; J. Hazard. Mater. 2016, 304, 417.
• ³¹
  da Silva, M. C. H.; da Silva, L. H. M.; Paggioli, F. J.; Coimbra, J. S. R.; Minim, L. A.; Quim. Nova 2006, 29, 1332.
• ³²
  Campos, R. A.; Patricio, P. R.; Vargas, S. J. R.; da Silva, L. H. M.; Hespanhol, M. C.; An. Acad. Bras. Cienc. 2018, 90, 1929.
• ³³
  Bulgariu, L.; Bulgariu, D.; Stud. Univ. Babes-Bolyai, Chem. 2009, 4, 273.
• ³⁴
  Christensen, A. N.; Lebech, B.; Dalton Trans. 2012, 41, 1971.
• ³⁵
  Miersch, L.; Rüffer, T.; Schlesinger, M.; Lang, H.; Mehring, M.; Inorg. Chem. 2012, 51, 9376.
• ³⁶
  da Silva, L. H. M.; Loh, W.; J. Phys. Chem. B 2000, 104, 10069.
• ³⁷
  Lange, N. A.; Dean, J. A.; Lange's Handbook of Chemistry; McGraw-Hill: New York, 1979.
• ³⁸
  Patricio, P. R.; Cunha, R. C.; Vargas, S. J. R.; Coelho, Y. L.; da Silva, L. H. M.; da Silva, M. C. H.; Sep. Purif. Technol. 2016, 158, 144.
• ³⁹
  Bulgariu, L.; Bulgariu, D.; Sep. Purif. Technol. 2013, 118, 209.
• ⁴⁰
  Lothenbach, B.; Ochs, M.; Wanner, H.; Yui, M.; JNC-TN--8400-99-011: Thermodynamic Data for the Speciation and Solubility of Pd, Pb, Sn, Sb, Nb and Bi in Aqueous Solution; Japan Nuclear Cycle Development Institute: Naka-gun, Ibaraki, Japan, 1999.
• ⁴¹
  Sadler, P. J.; Muncie, C.; Shipman, M. A. In Biological Inorganic Chemistry, Structure and Reactivity, 1^st ed.; Bertini, I.; Gray, H. B.; Stiefel, E. I.; Valentine, J. S., eds.; University Science Book: Sausalito, California, 2007, p. 95.
• ⁴²
  Ojebuoboh, F. K.; JOM 1992, 44, 46.

Publication Dates

• Publication in this collection
  24 Oct 2019
• Date of issue
  Nov 2019

History

• Received
  1 Feb 2019
• Accepted
  28 June 2019