Liquid-Liquid Extraction of Ferric Ions into the Ionic Liquids

Abstract

Imidazolium ionic liquids containing acetylacetone, thenoyltrifluoroacetone, or 8-hydroxyquinoline, respectively, were used as the extracting agents for the separation of traces of iron (III) from its aqueous solutions with or without citric and oxalic acids. The results show that 8-hydroxyquinoline in imidazolium ionic liquids extract iron quantitatively from all the tested solutions including complexing ones, regardless of indications of unexpected iron behavior/speciation.

1. Introduction

At present days, ionic liquids become increasingly interesting for researchers in a wide range of fields. They play an important role in many current technologies from hydrometallurgy to food processing [1]. These substances gained the attention of scientists due to their unique and tunable properties depending on the exact structure of the anion and cation [2]. Among their advantages, wide liquid range, high thermal stability, low melting points, and a near absence of vapor pressure may be listed, too [3].

Although the possibility of tuning the properties of ionic liquids can lead to their specific use, ionic liquids can also be used as a substitute for conventional organic solvents, when they may be used in nearly any area where the conventional molecular solvents are traditionally employed. Liquid-liquid extraction is shown to be one of the most evident applications to purify or concentrate the substances [1] and it also seems to be a very promising way for the separation of metals and radionuclides in the treatment of industrial waste.

The imidazolium ionic liquids are usually used as solvents for suitable extracting agents and they have been successfully tested in the extraction of a variety of metals [1,4]; in many cases, the combination with extracting agent caused the synergistic effect [5]. Additional advantages of this group of ionic liquids are their acceptable viscosity of the ones with short carbon chains, and relatively high (electro)chemical stability, which allows various possibilities for their regeneration [5,6,7,8].

Extraction of iron from high to trace concentrations plays important role in many of the fields mentioned above, i.e., waste streams treatment, determination of iron in minerals, separation, and purification of radionuclide solutions, in water treatment, organic synthesis, and many others [3,5].

In the literature, the possibility of liquid-liquid extraction of ferric ions using different types of ionic liquids is mentioned, too. E.g., [9] authors studied solvent ex-traction using tricaprylmethylammonium thiosalicylate [A336][TS] and trihexyl(tetradecyl)phosphonium thiosalicylate, [PR4][TS], a thiol-containing task-specific ionic liquids for the removal of iron (III) from hydrochloric acid solutions. In these cases, Fe(III) ions were effectively removed from the aqueous phase by the phosphonium and ammonium ionic liquids and extraction efficiency increased along with hydrochloric acid concentration. The extraction was related to negatively charged iron (III) chloride complexes and its disadvantage is just the required higher HCl concentration.

In the other work [10] Matsumiya et al. focused on removal of iron by solvent extraction using tri-n-butyl phosphate (TBP)/triethyl-pentyl-phosphonium (P2225) bis(trifluoromethyl-sulfonyl)amide (TFSA) system. They found out that nine extraction cycles led to the removal of most of the Fe.

Hayati N. et al. [11] developed an iron liquid-liquid extraction method for groundwater treatment using a room-temperature ionic liquid, 1-butyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide; chelating agent 1,10-phenanthroline was used as the extractant. In that case, extraction of Fe(III) and Fe(II) was influenced significantly by the pH of the aqueous phase, and successful removal of iron was achieved. However, the authors state that detailed research is needed before the ionic liquid method can replace the conventional groundwater treatment protocol because the recovery rate was very low upon reuse.

The main aim of the research was to study the extraction of Fe(III) in trace concentrations into the selected ionic liquids with various extracting agents, from nitric acid aqueous solution, and in the presence of complexing agents in the aqueous phase.

2. Materials and Methods

2.1. Chemicals

Five ionic liquids, listed below, were purchased from Iolitec, Germany (high-purity grade, 99%) and used without additional purification.

• 1-ethyl-3-methylimidazolium bis (trifluoromethanesulfonyl) imide-[C₂mim][NTf₂]
• 1-butyl-3-methylimidazolium bis (trifluoromethanesulfonyl) imide-[C₄mim][NTf₂]
• 1-hexyl-3-methylimidazolium bis (trifluoromethanesulfonyl) imide-[C₆mim][NTf₂]
• Methyltrioctylammonium bis (trifluoromethanesulfonyl) imide-[N₁₈₈₈][NTf₂]
• Tributylmethylammonium bis (trifluoromethanesulfonyl) imide-[N₁₄₄₄][NTf₂]

For comparison with classic volatile organic compounds (VOC) representatives, benzene (≥99.7%, Lachema, Brno) and chloroform (pure, Lachema, Brno) were used.

As the extraction agent, acetylacetone (AcAc, purity ≥98%, Carl Roth GmbH + Co. KG, Karlsruhe, Germany), 8-hydroxyquinoline (8HQ, SIGMA-ALDRICH CHEMIE GmbH 99%, St. Louis, USA) and thenoyltrifluoroacetone (TTA, purity 99%, SIGMA-ALDRICH CHEMIE GmbH, St. Louis, USA) were tested.

Nitric acid, hydrochloric acid, sodium tetraborate decahydrate, and citric acid (H₃Cit) (Lach-Ner s.r.o., Neratovice, p.a., Czech Republic), sodium dihydrogenphosphate, sodium hydroxide (Penta, Prague, p.a.), and sodium hydrogenphosphate, and oxalic acid (H₂Ox) (Lachema Brno, p.a.) were used for the preparation of aqueous phase.

The aqueous phase was spiked with a stock solution of ⁵⁹Fe prepared from commercially available radionuclide solution of carrier-free ⁵⁹FeCl₃ in 0.5 M HCl, with reference activity A = 37 MBq (99%, Perkin Elmer, Waltham, MA, USA).

2.2. Instrumentation and Methods

The organic phase was prepared by dissolving an appropriate amount of extracting agent in respective ionic liquid or benzene or chloroform.

The aqueous phase varied according to the needs of the study:

• HNO₃ in a concentration of 0.001–1.5 mol L⁻¹ or
• 0.2 M phosphate buffer or
• 0.1 M borate buffer or
• 0.01 M H₂Ox or 0.005 M H₃Cit or the mixture of 0.01 M H₂Ox and 0.005 M H₃Cit

In all cases, the distribution ratio (D) and the percent of extraction (E) were calculated from the net count rates of the analyzed nuclide in the aqueous and organic phases measured at the same conditions, Equation (1).

$$\mathrm{D}=\frac{c_{org}}{c_{aq}}\approx \frac{I_{org}}{I_{aq}}, \mathrm{E}=100 \frac{c_{org}}{c_{tot}}\approx 100\frac{I_{org}}{I_{aq}+I_{org}}$$

(1)

where c_org, c_aq are concentrations and I_org, I_aq net specific count rates in the aqueous and organic phases, respectively, and c_tot total concentration of the analyte. The maximum and minimum determinable distribution ratios (D_max, D_min) and percent of extraction (E_max, E_min) were calculated using count rate limits of detection for each experimental set [12].

The uncertainty of the distribution ratios was evaluated as a combined uncertainty of the statistics of the measurement and pipetting and it did not exceed 5% in all experiments.

Extraction experiments were performed as follows:

• Equal volumes (1 mL) of organic and aqueous phases were contacted in Eppendorf microvials and shaken for 30 min at laboratory temperature.
• Aqueous and organic phases were separated by centrifugation (5 min at 700 RCF)
• 200 μL aliquots were taken from both phases and measured in a well-type 2″ NaI (Tl) scintillation detector connected to an NV 3102 single-channel counter.
• After the separation of phases, the equilibrium pH values of the aqueous phase were measured using a digital pH meter PHM240 equipped with a combination Red-Rod glass electrode.

3. Results and Discussions

Pre-experimental screening confirmed the presumption of negligible iron (III) extraction into pure imidazolium ionic liquid. Therefore, the selection of extractants based on their solubility in the tested ionic liquids, and their extraction abilities was performed. Based on the results, acetylacetone (AcAc), 8-hydroxyquinoline (8HQ), and thenoyltrifluoroacetone (TTA) were selected for detailed extraction experiments including parameters strongly influencing the extraction, such as pH of an aqueous phase, different types of ionic liquids, the composition of the organic phase, concentration of extractant, and presence of organic complexing agents in an aqueous phase, were studied in detail.

3.1. Acetylacetone

As it is shown in Figure 1a, the trend of the extraction dependence on pH value with acetylacetone as the extracting agent (pKa = 8.82) [10] is similar for all used solvents. The highest extraction was achieved in the pH range 4–9, and around pH 11, the drop of the E was observed. In the pH range with maximal extraction, the E of Fe(III) extraction into acetylacetone in different solvents increased in the order of [N₁₄₄₄][NTf₂] < [N₁₈₈₈][NTf₂] < [C_xmim][NTf₂] < benzene. In this range, more than 98% of iron was extracted into AcAc solution in [C₄mim][NTf₂]. The efficiency of extraction into ammonium-based ionic liquids ([N₁₄₄₄][NTf₂] and [N₁₈₈₈][NTf₂]) is lower—the achieved maximum extraction was no more than 85%. Compared to benzene as the conventional VOC mainly used for dissolving acetylacetone where almost quantitative extraction was reached, the Fe(III) extraction into selected ionic liquids is less effective.

The distribution ratio (D) dependence on equilibrium pH (Figure 1b) showed strictly linear behavior in the pH range 0–4. Slopes of these linear trends and related pH_1/2 values (pH at equilibrium at which D = 1) were calculated (

Table 1. Slopes and pH_1/2 values calculated from the dependence of D on equilibrium pH for extraction of ferric ion into 0.5 M AcAc in various ionic liquids or in benzene. (V_org/V_aq = 1, T = 25 °C, t(contact) = 30 min).

	[C2mim][NTf2]	[C4mim][NTf2]	[C6mim][NTf2]	[N1444][NTf2]	[N1888][NTf2]	Benzene
Slope	0.93 ± 0.03	1.19 ± 0.60	1.33 ± 0.08	1.17 ± 0.03	1.20 ± 0.17	1.64 ± 0.05
pH1/2	1.31 ± 0.06	1.46 ± 0.10	1.59 ± 0.14	2.33 ± 0.08	2.15 ± 0.31	1.06 ± 0.05

). The highest slope value was found for the benzene environment, in the case of ILs the slope value as well as the pH_1/2 increases with the alkyl chain length of imidazolium ionic liquids (C_xmim). In the case of ammonium ionic liquids, the possible trend is hidden within the uncertainty interval.

In addition, the dependence of distribution ratio (D) on the concentration of extracting agent in ionic liquid at constant pH was compared to the same dependence for benzene as a solvent (Figure 2) and it shows that increasing AcAc concentration in the [C₄mim][NTf₂] system is less effective when compared to benzene system; ionic liquid influences chelate behavior and related extraction mechanism.

3.2. Thenoyltrifluoroacetone

For this extraction agent (pK_a = 6.23) [13] only benzene as conventional VOC solvent and [C₄mim][NTf₂] representing ionic liquids were tested. As in the case of AcAc, the trend of the Fe(III) extraction dependences on equilibrium pH for IL and benzene as solvents are similar. The maximum extraction efficiency (98%) was reached at pH close to 4 (Figure 3a).

Distribution ratio dependence on equilibrium pH showed several linear sections in logarithmic scale in the pH range 0–4 indicating a change of the extraction mechanism (Figure 3b). The outer sections are quite short to characterize, however the middle section allowed to calculate slopes and appropriate pH_1/2 values, all summarized in

Table 2. Slopes and pH_1/2 values calculated from the dependence of D on equilibrium pH for extraction of ferric ion into 0.1 M TTA in [C₄mim][NTf₂] or in benzene (V_org/V_aq = 1, T = 25 °C, t(contact) = 30 min).

	[C4mim][NTf2]	Benzene
slope	1.79 ± 0.24	1.82 ± 0.12
pH1/2	2.78 ± 0.41	1.91 ± 0.19

. It can be seen that the slope achieved for [C₄mim][NTf₂] is steeper than that achieved for benzene as a solvent, but in the tested concentration range, TTA in [C₄mim][NTf₂] provides lower distribution ratios. However, the curve line describing the extraction of Fe(III) into IL and appropriate pH_1/2 value is shifted to higher pH values compared to the extraction into organic phase with benzene.

In addition, the dependences of distribution ratio on the concentration of TTA were tested at pH = 2.3 for both solvents (Figure 4). The small difference between the slopes can be seen indicating again different behavior of the chelate in the case of ionic liquid solvent.

3.3. 8-Hydroxyquinoline

The effect of different ionic liquids on the extraction of Fe(III) using amphoteric chelating agent 8-hydroxychinoline (8HQ) as extracting agent was studied. It is an amphoteric chelating agent with hydration constant pK_H = 5.0 and dissociation constant pKa = 9.66 [14]. As the reference organic solvent, chloroform was used in this case because it is commonly used as the most suitable VOC solvent for 8-hydroxyquinoline [6]. As in both cases above, the trend of the dependences on the equilibrium pH is similar for all tested solvents (Figure 5a)—imidazolium ionic liquids, ammonium ionic liquids, and chloroform. At pH ~ 1, the extraction of iron rises sharply and the maximal extraction (>99.7%) is achieved at pH ~ 3 for all solvents. Using ionic liquids, the extraction did not decrease significantly in the studied pH range (up to 9). In chloroform, a slight reduction of extraction at pH > 6 was observed.

The semi-logarithmic dependence of distribution ratio on the equilibrium pH value shows two easily definable sections with different slopes (Figure 5b), which suggests that the reaction mechanism is most likely to change. The parameters characterizing the log D dependence on pH value are summarized in

Table 3. Slopes and pH_1/2 values calculated from the dependence of D on equilibrium pH for extraction of ferric ion into 0.1 M 8HQ in various ionic liquids or in chloroform (V_org/V_aq = 1, T = 25 °C, t(contact) = 30 min). Bracketed values are approximations due to the limited number of experimental values.).

	[C2mim][NTf2]	[C4mim][NTf2]	[C6mim][NTf2]	[N1444][NTf2]	[N1888][NTf2]	CHCl3
Slope pH ~ 1.2	3.19 ± 0.26	3.82 ± 0.15	3.38 ± 0.60	2.80 ± 0.19	(2.97)	4.96 ± 0.12
Slope pH ~ 2	(0.66)	1.00 ± 0.06	1.24 ± 0.19	1.48 ± 0.18	1.44 ± 0.03	0.66 ± 0.03
pH1/2	1.66 ± 0.19	1.37 ± 0.08	1.35 ± 0.34	1.49 ± 0.13	1.53 ± 0.05	1.19 ± 0.04

. The effect of alkyl chain length cannot be quantified as well as the trend in pH_1/2 values, in all the cases the differences do not exceed the uncertainty intervals. However, for all ionic liquids systems and at the given 8-HQ concentration, the pH_1/2 values are higher than in the reference CHCl₃ system.

The distribution ratio dependences on various 8-HQ concentrations in [C₄mim][NTf₂] in logarithmic scales at two different equilibrium pH values were measured (Figure 6). From these data, the slopes of these dependencies were calculated. At lower pH, the slope is between 1.5 and 2 (1.78 ± 0.12), but at a higher pH value, it is close to one (0.95 ± 0.03). This indicates a change in the number of 8-HQ in the extracted species above pH ~ 2.

3.4. Extraction from Aqueous Solution Containing Common Complexing Agents

As the representatives of the substances widely known as complicating the iron extraction, namely in trace concentrations, citric acid, and oxalic acid were selected. All experiments were performed with a solution of single acids or their mixture at the natural initial pH of the solutions (pH = 2.1–3). Based on the results above, [C₄mim][NTf₂] was used as a representative of ionic liquids and for comparison, the respective classical solvents were used.

All experimental results are summarized in

Table 4. The effect of citric and oxalic acids on the extraction of ferric ion into various organic phase. (V_org/V_aq = 1, T = 25 °C, t(contact) = 30 min).

	No Complexant	H3Cit		H2Ox		0.005 M H3Cit + 0.01 M H2Ox
		0.005 M	0.01 M	0.005 M	0.01 M
0.5 M AcAc in benzene	97.5 ± 0.5%	98.9 ± 0.4%	95.2 ± 0.4%	2.3 ± 0.1%	5.2 ± 0.1%	5.1 ± 0.1%
pHeq	2.3	2.8	2.6	2.4	2.1	2.1
0.5 M AcAc in [C4mim][NTf2]	94.7 ± 0.5%	91.5 ± 0.4%	85.8 ± 0.4%	0.4 ± 0.1%	0.1 ± 0.1%	0.1 ± 0.1%
pHeq	2.7	2.8	2.6	2.4	2.1	2.1
0.1 M TTA in [C4mim][NTf2]	33.8 ± 0.2%	3.3 ± 0.1%	0.2 ± 0.1%	4.3 ± 0.1%	0.4 ± 0.1%	0.2 ± 0.1%
pHeq	2.3	2.6	2.4	2.8	2.0	2.0
0.4 M TTA in [C4mim][NTf2]	87.0 ± 1.1%	15.1 ± 0.1%	2.9 ± 0.1%	14.9 ± 0.1%	1.1 ± 0.1%	1.5 ± 0.1%
pHeq	3.3	2.6	2.4	2.8	2.1	2.1
0.1 M 8HQ in CHCl3	>99.7%	>99.7%	>99.7%	22.3 ± 1.3%	6.7 ± 1.2%	5.9 ± 0.1%
pHeq	3.1	3.8	3.5	3.8	3.4	3.3
0.1 M 8HQ in [C4mim][NTf2]	>99.7%	>99.7%	>99.7%	>99.7%	87.2 ± 0.4%	74.2 ± 0.3%
pHeq	4.3	5.0	4.7	5.0	4.5	4.3
0.5 M 8HQ in CHCl3	>99.7%	>99.7%	>99.7%	97.9 ± 0.5%	79.8 ± 0.4%	82.0 ± 0.4%
pHeq	4.0	4.2	4.0	4.1	3.9	3.8
0.5 M 8HQ in [C4mim][NTf2]	>99.7%	>99.7%	>99.7%	>99.7%	>99.7%	>99.7%
pHeq	4.3	5.6	5.3	5.7	5.3	5.1

. It can be seen that the effect of citric acid-suppressing iron extraction is much smaller than that of oxalic acid in all tested systems. And—as supposed—with increasing concentration of complexing acid less iron is extracted. Any synergy or antergy was not observed between the complexing acids in the range of E uncertainties.

AcAc was shown to be strong enough to extract Fe(III) from the solution of citric acid. Compared to the extraction using benzene as the solvent, the extraction was up to 10% lower with respect to the concentration of citric acid. In the case of oxalic acid and the mixture, the extraction did not exceed 6%. In this system, the use of ionic liquids and combination with AcAc is not bringing any advantage compared to the AcAc solution in benzene; and AcAc is in both cases not strong enough to pass complexing abilities of citric and namely oxalic acid.

Although iron is extracted relatively efficiently from HNO₃ with TTA, its extraction ability when using oxalic and/or citric acid solutions is relatively small. Even an increase of TTA concentration did not result in effective Fe(III) extraction.

The extraction with 8-HQ was much more efficient than with acetylacetone and TTA. At an extractant concentration of 0.5 mol.L⁻¹, complete extraction was achieved from all solutions, including oxalic acid solutions and a mixture of both complexing acids. Additionally, it can be clearly seen that the extraction into ionic liquid is much more efficient than into CH₃Cl.

4. Conclusions

Our experiments and results covered the area of the chelate-assisted extraction of iron (III) into ionic liquids. With respect to the aims of the work, most of the measured values are suitable for further investigation and deeper study on water/ionic liquid biphasic systems.

The achieved results clearly show that iron (III) is extracted from nitric acid solutions and from a given aqueous complexing environment. From the set of chelating agents and regarding organic phase composition, iron is quantitatively extracted with 0.5 M 8-hydroxyquiniline solution in [C₄mim][NTf₂] at the natural pH of the solution. Such a result indicates the ability of such a system to separate iron and concentrate it from diluted systems without any additions of buffers or mineral acids, which can further influence the speciation and purity of the resulting separated product.