Relationship Between Extraction of Arsenic via Ion Solvation and Hansen Solubility Parameters of Extractants

ABSTRACT Extraction of As(III) from hydrochloric acid solution using various solvents on the basis of ion solvation has been reported, but a theoretical framework to describe the suitability of solvents for extraction has not been systematized. In this study, comprehensive extraction tests for As(III) using a variety of organic solvents were conducted to clarify the requirements for solvents to extract As(III). From the results of initial screening tests, various aromatics, ethers, and ketones were rated as candidates for the extraction of As(III) at high HCl concentrations, whereas aliphatic hydrocarbons were excluded. Many solvents showed high extraction capacity for As(III) to give organic fraction concentrations of 150 mM or higher. The logP value (P = partitioning coefficient between n-octanol and water) within a solvent class showed some correlation with extractability of As(III) but the correlation was invalid across solvent types. High correlation was shown between the Hansen solubility parameters of the solvents and As(III) extractability: Twenty of 22 solvents (90.9%) were appropriately classified as valid or invalid for the extraction of As(III) based on the Hansen sphere. 2-Nonanone was recommended as an ideal solvent for As(III) extraction because of its good balance between high extractability and its desirable physical properties for industrial operations. As(III) was quantitatively stripped from 2-nonanone into water with a concentration factor of 22.6.


Introduction
Arsenic is known as a contaminant that shows extremely high toxicity in humans. [1,2]As a result, the removal of arsenic from process streams is an important subject. [3,4]Arsenic is associated with metal deposits as arsenopyrite (FeAsS) and arsenian pyrite [Fe(As,S)2]. [5,6]Gold is present in arsenic-and carbon-containing, high-sulfur ores. [7]Arsenic is also present in the copper smelter flue dusts that are trapped and collected by dust collectors. [8]fter oxidative pretreatment in hydrometallurgical process, arsenic is dissolved in aqueous media as ionic species.While separation processes such as coprecipitation and adsorption can be used to remove arsenic, this study focuses on the use of solvent extraction of As(III) from hydrochloric acid media based on ion solvation extraction.As(III) can be extracted from highly concentrated hydrochloric acid solutions using various organic solvents.Extraction of As(III) has been reported since the 1950s, with examples including extraction with bis-(2-chloroethyl)ether, isopropyl ether, benzene, chloroform, or carbon tetrachloride from highly concentrated hydrochloric acid (>8 M) solution. [9,10]Sella and coworkers [11] gave a detailed report on the extraction of As(III) from hydrochloric acid solutions using aromatic solvent Solvesso 150.Addition of calcium chloride increases the activity of hydrochloric acid, resulting in increased extraction of As(III) with the solvent.Moreover, solubilization with 9 M HCl and subsequent extraction with chloroform has been used as a method for the selective quantitative determination of inorganic arsenic in seafood. [12]rganic solvents containing oxygen atoms such as ethers or ketones can be used to extract As(III).Dibutyl carbitol (DBC) is a commercially available extractant that can be used to extract As(III) based on ion solvation. [13]Methyl isobutyl ketone (MIBK), a ketone compound, also shows extraction ability for As(III). [14,15]Recently, the authors reported extraction of arsenic, selenium, and antimony using cyclopentyl methyl ether (CPME). [16]CPME is a recently commercialized ethereal compound that is more hydrophobic than typical ethereal compounds such as diethyl ether or tetrahydrofuran.CPME extracted As(III) at high HCl concentration, and the back extraction was also quantitative.However, CPME has the disadvantages of relatively high aqueous solubility (11.0 g/L) and low flash point (−1°C), which make it unsuitable for use as an industrial solvent extractant.
As described above, extraction of As(III) from hydrochloric acid systems using several organic solvents has been reported.However, the structural requirements of an optimal solvent for the extraction of As(III) have not yet been clarified.In addition, there are also restrictions on the physical characteristics of solvents for use in industrial applications in terms of safety, continuous use, and operability.In the present study, tests of As(III) extraction were conducted using various organic compounds to examine the requirements for an organic solvent suitable for the extraction.In addition, the relationship between As(III) extractability using organic solvents and Hansen solubility parameters (HSPs) was considered, [17,18] because the solubility of the chloride complex in an organic solvent should influence the extractability.[21][22][23] In these precedents, the effect of diluents on the extraction of metal ions by the complexation with extractants such as chelating agents was investigated.On the other hand, the present study targets ion solvation extraction system in which there is no extractant other than solvent.Therefore so the effect of the solvent should be different from that using extractants.Additionally, previous studies have used the Hildebrand solubility parameter as an index.There should not be any other studies than our group that investigated the effect of Hansen solubility parameters, which consist of three components, on the extraction of metal ions. [24]urthermore, solvent suitability for As(III) extraction was assessed in consideration of the requirements for industrial use.Conditions for back extraction, metal separation, and concentration of As(III) using the selected solvent were also investigated.

Reagents
Analytical-grade sodium dioxoarsenate was purchased from Fujifilm Wako Pure Chemical Industries (Osaka, Japan).All other inorganic reagents were of analytical grade and were used as received.
[27] The names, abbreviations, and structures of organic solvents used in this study is summarized in Table 1.
The logarithm of the partitioning coefficient between n-octanol and water (logP) is used as a general index of the hydrophilic -lipophilic balance of a compound.The logP values of the extractants were estimated using MarvinSketch 6.2.1 software (ChemAxon, Budapest, Hungary) based on the KLOP method. [28]The HSPs of organic solvents were also used as indexes of As(III) extractability.The HSP values of solvents used in extraction tests were obtained from HSPiP (Hansen Solubility Parameters in Practice) software (ver.5). [17]

Liquid -liquid extraction tests
Liquid -liquid extraction tests for As(III) were conducted in batch mode as follows. [16,26,29,30]An aqueous solution of As(III) was prepared by dissolving sodium dioxoarsenate to form 1.0 × 10 −4 M solution in hydrochloric acid.The initial concentration of the hydrochloric acid solution was adjusted to an appropriate level between 0.10 and 8.0 M. Various organic solvents were used as extracting phases.The organic solvent (1.0 mL) and the aqueous solution (5.0 mL) were mixed in a glass sample bottle; the volume ratio between organic phase and the aqueous phase was fixed at 1:5 (O/A = 0.2).The mixture was shaken (120 rpm) in a water bath at 30°C for 1 h.After phase separation, the concentrations of As(III) in the aqueous solutions before and after extraction were determined by atomic absorption spectrophotometry (AAS; AA-7000, Shimadzu, Kyoto, Japan).The extraction percentage of As(III) was calculated according to Equation ( 1): where [As(III)] aq,init and [As(III)] aq,eq represent the initial and equilibrium concentrations of As(III) in the aqueous phase, respectively, and [As(III)] org,eq is the total concentration of the metal ion in the organic phase at equilibrium.Liquid -liquid extraction tests for Sb(III), Sb(V), and Se(IV) were also conducted as described for As(III).Furthermore, the extraction capacity of As(III) into organic solvents was investigated by contacting an organic solvent (1.0 mL) and an aqueous solution (5.0 mL) containing As(III) (1.0 × 10 −3 to 6.0 × 10 −3 M) in HCl (8 M).

Back extraction of As(III)
The forward extraction of 1.0 × 10 −3 M As(III) using 2-nonanone from 8.0 M hydrochloric acid was performed by contacting aqueous and organic phases (150 mL/30 mL, O/A = 0.2) in a similar manner as described above.The organic phase containing As(III) was divided into 1.0-mL portions, and each portion was contacted with a 5.0-mL portion of distilled water, 0.1-5.0M hydrochloric acid, 1.0 M nitric acid, or 0.50 M sulfuric acid.Both phases were mixed and shaken at 30°C for 24 h.The stripping solution was separated from the organic phase and the back extraction (B.E.[%]) was calculated according to Equation ( 2): where [As(III)] org,init represents the initial concentration of extracted species in the organic phase and [As(III)] aq,strip is the total concentration of the extracted species in the aqueous phase after stripping.

Enrichment of As(III)
The forward extraction of 1.0 × 10 −3 M As(III) using 2-nonanone from 8.0 M hydrochloric acid was performed by contacting aqueous and organic phases (200 mL/40 mL, O/A = 0.2) in a similar manner as described above.After extraction, the organic phase containing As(III) was divided into 2.0-mL portions.The portions were contacted with 0.40, 1.0, or 2.0 mL of distilled water.After shaking the mixture at 30°C for 24 h, the aqueous solution was separated from the organic phase and the concentration of As(III) was determined by AAS.The enrichment ratio of As(III) was calculated as the ratio of the concentration of As(III) in the stripping solution to the concentration of the initial aqueous solution.

Extraction of As(III) using various organic solvents
Generally, when As(III) is extracted from concentrated HCl solution, the predominant As(III) species in solution is AsCl 3 . [11]Therefore, experiments investigating the extraction of As(III) using different solvents were first performed at different HCl concentrations.Extraction based on ionic solvation is generally very rapid, [16,26,[29][30][31] with As(III) extraction using 2-NON reaching equilibrium within 20 min (Fig. S1).
Figure 1 shows extraction profiles for As(III) using various organic solvents as a function of hydrochloric acid concentration.As shown in Figure 1a, extraction with aliphatic hydrocarbons was low regardless of the length of the carbon chains.Aromatic hydrocarbons are known to act as extractants for As(III), [32] and the order of extraction ability (Figure 1b) was observed to decrease with increasing length of alkyl side chain: toluene, ethylbenzene > butylbenzene > nonylbenzene, dodecylbenzene.Aliphatic ethers and aromatic ethers showed high extractability from 8 M HCl solution.The extraction percentages using CPME (63.7%) and DBC (60.5%) were comparable (Figure 1c) and higher than that of dibutylether (42.8%).The results in Figure 1d suggest two tendencies in the extraction using aromatic ethers.
The data shown in Figure 1 indicate that various aromatics, ethers, and ketones have good extractability for As(III) at high HCl concentration.In addition, the length of the alkyl chain and the number of functional groups such as ethereal oxygen also influence the extractability.These dominant factors for As(III) extraction were studied in subsequent experiments.

Extraction capacity for As(III)
Figure 2 shows the extraction behavior of As(III) using various organic solvents as a function of initial As(III) concentration in 8 M HCl solution.Considering that the volume ratio between the organic phase and the aqueous phase was fixed at 1:5 (O/A = 0.2), As(III) was concentrated fivefold in the extracting organic phase in the case of quantitative extraction.The concentration of As(III) in the organic phase increased with increased initial concentration of As(III) and was not saturated.The order of the extraction capacities using each solvent showed the same tendency as that of the extractability study: (i) aliphatic hydrocarbon (dodecane) was unsuitable for As(III) extraction; (ii) order of extraction capacities using aromatic hydrocarbons, aromatic monoethers, aromatic diethers, and aliphatic ketones decreased with the increased length of alkyl side chain.Nine solvents (toluene; CPME; DBC; MB; 1,2-DMB; 1,2-MOB; 2-OCT; 2-NON; 2-DOD) were able to be loaded to at least 150 mM As(III).][31] Figure 2 can also be plotted as extraction isotherms as a function of the equilibrium concentration of As(III) in the aqueous phase, which is shown as Fig. S2.

Requirements for organic solvents to extract As(III)
The alkyl chain length of solvent was observed to influence the extractability of As(III), suggesting that solvent hydrophobicity was a dominant factor for the extraction.Therefore, the relationship between extraction percentage of As(III) in 8.0 M HCl and the logP value of each solvent was analyzed as an indicator of the effect of hydrophilic -lipophilic balance (Figure 3 and Table 1).Within the solvent type, the extraction percentage decreased as the logP value of the solvent increased, suggesting that solvents with high The correlation between the HSPs of solvents and the extractability of As(III) was investigated to understand the importance of solvent properties in relation to extraction.In terms of HSPs, the solubility parameter is divided into the three parameters of nonpolar (dispersion) interactions (δ D ), polar (dipole -dipole and dipole-induced dipole) interactions (δ P ), and hydrogen bonding interactions (δ H ). [17,18,33] The three parameters can be visualized in a three-dimensional (3D) diagram.The solubility parameter distances between two materials (R a ) can be given by Equation ( 3): where δ D1 , δ P1 , and δ H1 represent the HSP values of a compound, while δ D2 , δ P2 , and δ H2 represent the HSP values of the other compound.A small R a value means that the HSP values of the compounds are close to each other.HSPiP software (ver.5) was used to estimate ideal HSPs for As(III) extraction.The six solvents (CPME; 1,2-DMB; MIBK; 2-OCT; 2-NON; acetophenone) having As(III) extractions of more than 50% in 8 M HCl were classified as "Valid for As(III) extraction (EX)", whereas 16 solvents (n-hexane; n-heptane; n-octane; 2,2,4-trimethylpentane; n-nonane; n-decane; n-dodecane; cyclohexane; toluene; o-xylene; p-xylene; ethylbenzene; n-butylbenzene; MB; diphenylether; dibenzylether) having As(III) extractions of less than 50% were classified as "Invalid for As (III) EX".The HSPs of 22 solvents were plotted in a three dimensional space to obtain the Hansen sphere using the classification based on As(III) extractability (Fig. S2).In the software HSPiP, the Hansen sphere was obtained by the classic binary fitting using a set of 0 (invalid) and 1 (valid) values to find a best fit based on the Hansen exponential penalty function for wrong in and wrong out. [17]The δ P and δ H values of valid solvents are relatively high compared with those of invalid solvents.The Hansen solubility sphere describes the 3D space within which the solvent HSPs are valid for As(III) extraction.The center coordinates of the Hansen solubility sphere are δ D = 18.43, δ P = 12.65, and δ H = 5.82 with a radius (R a ) of 9.2.HSPs of many polar solvents are located inside the solubility sphere.However, many of these solvents are too soluble in water for them to be effective in solvent extraction.Therefore, the solubility sphere was redefined to identify the valid solvents for extraction operations and for As(III) extraction.Thirty-four solvents with an aqueous solubility of more than 100 g/ L and water were classified as "Invalid for solvent extraction (SX)".The Hansen solubility sphere was regenerated using the classification based on As(III) extractability and validity for solvent extraction (Figure 4, Table S2).Extraction tests for As(III) using the solvents shown in Table S2 were not performed, and only their HSPs were used to narrow down the Hansen sphere.The center coordinates of the Hansen solubility sphere were δ D = 17.76, δ P = 7.69, and δ H = 2.33, and R a was 5.6.The recalculated solubility sphere is much smaller than the earlier version, suggesting that the range of HSPs for solvents to extract As(III) based on ionic solvation had been refined.Many solvents were properly classified at the interface of solubility spheres according to the extractability of As(III).
However, the HSPs of acetylacetone were located inside the sphere despite its lack of suitability for As(III) extraction because of high aqueous solubility, while the HSPs of 1,2-DMB were outside the sphere (R a = 1.489) despite its high extractability.These conflicting results suggest that the extractability of As(III) cannot be completely explained in terms of HSPs.Nevertheless, 53 solvents except acetylacetone (Ra = 0.997 but Invalid for SX), MB (Ra = 0.983 but Invalid for As(III) EX) and 1,2-DMB (Ra = 1.489 but Valid for As(III) EX) were properly classified by the resulting sphere.The results suggests that the specialized HSPs are useful to estimate the extractability of As(III) by untested solvents.

Physical properties of organic solvents
Solvents used for liquid -liquid extraction must satisfy multiple criteria. [26,29]able 2 summarizes the physical properties of 10 solvents that showed high extractability in Figure 2. The aqueous solubility should be as low as possible to reduce solvent loss to the aqueous phase.The aqueous solubilities of ethereal compounds CPME, DBC, MB, and 1,2-DMB are higher than those of other solvents, which is disadvantageous in repeated extraction operations.The aqueous solubilities of aromatic diethers and aliphatic ketones decrease with increased length of alkyl side chains: 1,2-DMB >1,2-MOB; 2-OCT >2-NON >2-DOD.Aromatic diethers are undesirable because their densities are close to that of water, which can induce phase inversion.Viscosities of most solvents except for aromatic diethers are low and acceptable for use in liquid -liquid extraction.For safe use in industrial applications, solvent flash point should be high.Although toluene and CPME showed good extractability for As(III), they both have a low flash point.However, the flash points of DBC, aromatic diethers, and aliphatic ketones are acceptable for industrial operation (flash point of 1,2-MOB is unknown but should be higher than that of 1,2-DMB).
After comprehensive consideration of the above viewpoints, aliphatic ketones have potential as extractants for As(III) because of their low aqueous solubility, low viscosity, and relatively high flash point.The extraction behavior of 2-NON was then investigated in detail because of its good balance between high extractability and desirable physical properties.Although the extractability of 2-OCT was higher than that of 2-NON, the relatively low flash point of 2-OCT is undesirable in industrial operations.

Selectivity of As(III) extraction
Extraction of various metal ions using 2-NON from hydrochloric acid media was investigated in a previous study. [29]In the current study, an additional extraction test for As(III), Sb(III), Sb(V), and Se(IV) was a Data from Oshima et al. [25] b Estimated using ChemDraw ver.19.0 (PerkinElmer, MA, USA).c Estimated using MarvinSketch 6.2.1 software (ChemAxon, Budapest, Hungary) and the KLOP method.
conducted to investigate the extraction behaviors of these toxic ions.
Although arsenic can be present as As(V), the extraction test was not conducted for As(V) because it was not extracted using DBC, MIBK, or CPME in a previous study. [16]igure 5 shows the extraction profiles of As(III), Sb(III), Sb(V), Se(IV), and various other metal ions from HCl solution (1 or 5 M).As previously reported, extraction of Au(III) increases with the increase of HCl concentration.Fe(III) and Ga(III) also showed good extraction at high HCl concentration.These trivalent metal ions are present as tetrachloro-complexes and the resulting anionic species (AuCl 4 − , FeCl 4 − , GaCl 4 − ) should be extracted using 2-NON.In contrast, extraction of divalent metal ions was low.Extraction of Sb(III) and Sb(V) increased with increased HCl concentration.Sb(V) is present as a hexachlorocomplex at high HCl concentration [34] and the anionic chloride complex SbCl 6 − accompanied with a proton should be the species extracted.Extraction of Sb(III) was lower than that of Sb(V).Extraction of Se(IV) at high HCl concentration was very high.From these results, it is apparent that 2-NON can be used for the separation of arsenic, antimony, and selenium from many kinds of metal ions except for Au(III), Fe(III), and Ga(III).

Back extraction of As(III)
Table 3 shows the back extraction of As(III) from 2-NON using various aqueous solutions.As(III) was quantitatively recovered from 2-NON in all  cases.In the present study, the extraction percentage was calculated based on the mass balance from the decrease of As(III) concentration in the aqueous phase.Quantitative stripping also suggests the validity of the mass balance.Addition of mineral acids (0.1-5.0 M) did not influence the back extraction.In the forward extraction, As(III) was extracted as AsCl 3 (arsenic trichloride).In the back extraction step, arsenic trichloride was hydrolyzed by contacting the aqueous solution containing a lower concentration of HCl.The resulting arsenous acid or intermediate species is not stable in 2-NON and can be stripped to the aqueous solution.

Enrichment of As(III) by extraction using 2-nonanone
Extracts from the extraction can be enriched by stripping using a small volume of stripping solution.Table 4 shows the enrichment ratio of As(III) based on the extraction using 2-NON followed by stripping using distilled water.The percentage of the forward extraction of As(III) using 2-NON was 67.5%.After stripping using distilled water, the concentration of As(III) was enriched by 4.3 to 22.6 times (compared with the initial solution) depending on the volume of the stripping solution.In hydrometallurgical processing, arsenic coexists as a dilute impurity with the target metal.The results of this study show that As(III) can be enriched by extraction with 2-NON and then by stripping back into water.

Conclusions
The extraction of As(III) based on ion solvation using various organic solvents from hydrochloric acid media has been reported for a long time.However, the requirements of solvents for optimal extraction of As(III) have not been clarified.In this study, the correlation between the properties of solvents and As(III) extractability was analyzed from the results of As(III) extraction tests using various solvents.As(III) is not extracted with non-polar aliphatic hydrocarbons, whereas aromatics, ethers, and ketones show high extractability at high HCl concentration and have a large extraction capacity.It is difficult to estimate As(III) extractability based only on the hydrophilic -hydrophobic balance of the solvent.However, HSPs of solvents provide useful information to estimate As(III) extractability.Solvents should have moderate polarity and hydrogen bonding properties to extract As(III), while low polarity is desirable for solvents to be used in solvent extraction.As the result, the range of HSPs of solvents used for As(III) extraction has been specified.This finding can be used to predict As(III) extractability for untested solvents.For industrial operation, solvent choice for As(III) extraction is limited because of the balance between extractability and physical properties.The results of this study suggest that 2-NON has good potential as a solvent for selective extraction and concentration of As(III).

Figure 4 .
Figure 4. Three-dimensional Hansen solubility parameter (HSP) diagram for solvents based on As(III) extractability in 8.0 M HCl.Blue keys: solvents that are valid for As(III) extraction; red keys: solvents that are invalid for As(III) extraction because of low extractability or high aqueous solubility.

Table 1 .
The logP values of each organic solvent and the extraction percentage of As(III) in 8.0 mol/ dm 3 HCl.

Table 2 .
Physical properties of various organic solvents.