Green speciation of iron using aqueous two-phase system

Fe(II) and Fe(III) have distinct chemical and biological functions. Consequently, it is more important to determine the fraction of both oxidation state that knowing the total iron concentration in a sample. However, green methods for iron speciation are still limited. This work uses aqueous two-phase system, a safe alternative to liquid–liquid extraction, to perform the chemical speciation of iron. This method is based on the reaction of Fe(II) with 1,10-phenanthroline extractant, forming a complex of Fe(II)-phenanthroline that concentrates in the top phase of the system. The Fe(III) specie concentrated in the bottom phase of the system. Iron speciation was affected by the electrolyte nature, macromolecule type, quantity of phenanthroline added, and pH. The system formed by PEO1500 +Na3C6H5O7 +H2O at pH 6.00, containing 5.00 mmol kg of phenanthroline, was successfully used to separate the iron species before determination by flame atomic absorption spectrometry. Under these optimal conditions, a separation factor of 233 was obtained between Fe(II) and Fe(III) with extraction percentages of (95.1 ± 1.0)% and (7.68 ± 0.50)%, respectively The proposed method was successfully applied for iron speciation in water samples, and provided recovery percentages ranging between 90 and 106%.


INTRODUCTION
In nature, the same element can exist in different chemical forms (different isotopic compositions, oxidation state, or type of binding ligand), and each one can have a different toxicity, mobility, and bioavailability (Safavi et al. 2002, Kiss et al. 2017).Thus, it is more important to determine the concentration of each one rather than to determine the total concentration of an element in a sample.
Iron is an essential element in the biochemistry of organisms, since the simple ones as 1930 RAQUEL A. CAMPOS et al. microorganisms to complexes as humans (Vukosav et al. 2012).It has a central hole in chemical and biological functions as phytoplankton growth, respiratory, photosynthetic electron transport chains, synthesis of chlorophyll, and reduction of nitrate and nitrite (Lin andTwining 2012, Vukosav et al. 2012).Iron is generally found as Fe(II) and Fe(III) species and its crucial functions are dependent of the oxidation states (Pohl andPrusisz 2006, Suárez et al. 2015).Fe(II) is essential for the smooth transport and storage of oxygen in animals, while its oxidized form does not bind to oxygen (Safavi et al. 2002).Especially in an aquatic environment, it is important to identify the iron species, since its chemical form influences its bioavailability and the physicochemical properties of other elements and organic substances (Bulgariu and Bulgariu 2010).The oxidation state of iron in an environment can also indicate its electrical potential and microbial activity (Moghadam et al. 2011).Therefore, it is important to determine not only the total iron content, but also the fraction of Fe(II) and Fe(III) species.The tool to identify and/or measure the quantity of one or more individual chemical species in a sample is called "chemical speciation" according to the International Union of Pure and Applied Chemistry (IUPAC) (Templeton et al. 2000).
One of the most widespread analytical techniques for the determination of metals is flame atomic absorption spectrometry (FAAS) because of its cost and ease of operation (Paleologos et al. 2002, Tasev et al. 2006, Bahar and Zakerian 2012).However, FAAS cannot distinguish between the different chemical forms of an element.Therefore, it is necessary a sample preparation step, in which the species are separated, for chemical speciation using FAAS is performed (Tabrizi 2010).
An interesting strategy to separate iron species is a liquid-liquid extraction (LLE) process (Pournaghi-Azar and Fatemi 2000, Tasev et al. 2006, Pepper et al. 2010).This process is based on the distribution of the analyte between an aqueous phase and an organic phase containing an extractant, such as thiocyanate (Tasev et al. 2006), 1,10-phenanthroline (Tasev et al. 2006), bis (2-ethylhexyl) hydrogen phosphate (Pepper et al. 2010), or 8-hydroxyquinoline (Pournaghi-Azar and Fatemi 2000).A simple procedure, LLE allows the separation of species using different solvents (isobutylmethyl ketone (Tasev et al. 2006), heptane (Pepper et al. 2010), or chloroform (Pournaghi-Azar and Fatemi 2000).However, LLE involves the use of toxic, carcinogenic and/or flammable organic solvents, and typically requires a long time for phase separation, since emulsions can be formed.These disadvantages limit the applications of the solvent extraction (SE) technique, making it necessary to develop new separation procedures that are environment-friendly, safer, faster, and cost-effective.
An environmentally safe and promising alternative for the extraction and/or separation of metals is the use of an aqueous two-phase system (ATPS).The ATPS consists of two immiscible liquid phases, spontaneously formed by combining aqueous solutions of polymer-polymer (Persson et al. 2000), polymer-electrolyte (Patricio, et al. 2011a), ionic liquid-electrolyte (Bridges et al. 2007), or ionic liquid-polymer (Freire et al. 2012) under specific thermodynamic conditions of temperature, pressure, and composition (Freire et al. 2012).Each formed phase is rich in one of the components, but water is the major component in both phases (Martins et al. 2008, De Andrade et al. 2011, De Lemos et al. 2011, Patricio et al. 2011a, Virtuoso et al. 2012).This LLE system has a high separation efficiency and low cost; it is also easy to operate, does not require the use of organic solvents, and some components can be recycled.As a clean and economical alternative, it has been explored for the purification, separation, preconcentration, and/or determination of biomolecules (Reh et al. 2002, Haraguchi et al. 2004, Azevedo et al. 2009), phenolic compounds (Rodrigues et al. 2010, 2011), dyes (Chethana et al. 2007, Mageste et al. 2009, 2012), and metallic ions (Patricio et al. 2011b, De Lemos et al. 2012, Rodrigues et al. 2012, Da Cunha et al. 2016).However, few studies have explored the separation of species using ATPS, some of which include studies on mercury (Roy and Lahiri 2009), chromium (Akama and Sali 2002, Roy and Lahiri 2008, Patricio et al. 2016, Simonova et al. 2016), iodine (Griffin et al. 2004), and iron (Bulgariu and Bulgariu 2010).
Only one study proposes the use of ATPS to prepare samples for the separation of iron species.Bulgariu and Bulgariu (2010), used the poly(ethylene oxide) + (NH 4 ) 2 SO 4 + H 2 O ATPS to promote a selective extraction of Fe(III) in the presence of Fe(II).While Fe(III) was quantitatively extracted to a polymer-rich phase with the help of the thiocyanate extractant, the Fe(II) ions were concentrated in the (NH 4 ) 2 SO 4 -rich phase.The authors investigated the influence of pH, salt stock solution, and extractant concentration in the ATPS, on the separation of the species.However, they did not evaluate the influence of the system composition, the nature of the ATPS-forming polymer or electrolyte, or especially, the possible interconversion of the iron species in the system.
In the present work, a new green method is proposed for the chemical speciation of iron.This method involves the separation of Fe(III) and Fe(II) using ATPS, followed by determination of the species with the non-selective FAAS technique.While Fe(II) preferentially concentrated in the polymer-rich phase (top phase, TP), with the aid of the 1,10-phenanthroline (phen) extractant, the Fe(III) ions were partitioned to the salt-rich phase (bottom phase, BP).Parameters of the ATPS that could affect the separation of species, such as the nature of the macromolecule (L64 triblock copolymer or PEO1500 polymer), the use of environmentally safer electrolytes (Na 3 C 6 H 5 O 7 or Na 2 SO 4 ), the effect of a wide pH range, and the quantity of the extractant added in the system, were thoroughly investigated.The proposed method is green, simple, and allows the speciation of iron using FAAS.

EQUIPMENT
The pH measurements were carried out with a pH meter (Metrohm, 827, Herisau, Switzerland) equipped with a combined glass electrode.An analytical balance (Shimadzu, AY 220, São Paulo, Brazil) with an uncertainty of ± 0.0001 g was used to prepare the working solutions and the ATPS.The temperature of the ATPS was controlled by a thermostatic bath (Marconi, MA 184, São Paulo, Brazil) at (25.0 ± 0.1) °C.To accelerate phase separation, a centrifuge (Thermo Scientific, Heraeus Megafuge 11R, Osterode, Germany), was used.The concentration of the ions was determined by ultraviolet-visible molecular absorption spectrometry (Shimadzu, 2550, Kyoto, Japan) at 512 nm or by FAAS (Varian, AA240, Melbourne, Australia) using the following instrumental conditions: a wavelength of 248.3 nm, lamp current of 5.0 mA, spectral resolution of 0.2 nm, and flame composition (air/C 2 H 2 ) of 3.50/1.50L min -1 .

ATPS COMPOSITION
Stock solutions of the macromolecule and electrolyte were prepared using water at a pH of 1.00, 6.00, or 12.0, adjusted with H 2 SO 4 (5.37 mol kg -1 ) or NaOH (3.92 mol kg -1 ).Stock solutions of the macromolecules (L64 or PEO1500) and the electrolyte (Na 3 C 6 H 5 O 7 or Na 2 SO 4 ) were prepared in suitable concentrations to provide an ATPS with the compositions described in Table I, when these two solutions were mixed.
The ATPS used to extract the iron species was prepared by mixing 2.00 g of the working solution (extractant in the macromolecule solution) with 2.00 g of the electrolyte stock solution, prepared previous section, in a 10-mL centrifuge tube.This mixture was bubbled with nitrogen to prevent the oxidation of Fe(II) to Fe(III), following which 0.100 g of Fe(II) or Fe(III) stock solution was added to it, forming 0.100 mmol kg -1 of the iron specie in the ATPS.
Following the addition of the iron species, the centrifuge tube was manually stirred for 3 min, centrifuged at 10,000 rpm for 5 min at 25 °C, and left in a thermostatic bath at (25.0 ± 0.1)°C for 10 min to obtain two clear phases.The TP (or BP) aliquots were collected and appropriately diluted to determine the iron concentration by FAAS.All experiments were performed in triplicate.Eq. ( 1) was used to calculate the extraction percentage (% E): where n M x+ TP is the quantity of metal ions in the TP, n M x+ BP is the quantity of metal ions in the BP, and n M x+ Total is the total quantity of metal ions in the ATPS.

EXTRACTION OF Fe(II) IN THE PRESENCE OF Fe(III)
Fe(II) specie (5.64 mg kg -1 ) was added to the PEO1500 + Na 3 C 6 H 5 O 7 + H 2 O APTS at a pH of 6.00, which contained 5.00 mmol kg -1 of phen and different concentrations of Fe(III) (0.101, 0.507, or 1.01 mmol kg -1 ).Fe(II) was determined in the two phases of the ATPS by UV-vis spectrophotometry using the 1,10-phenanthroline method (Harvey et al. 1955, Eaton andFranson 2005).For this determination, aliquots of the two phases were collected and diluted in a phen aqueous solution (1.00 g kg -1 ), and the absorbances were measured at 512 nm.

DETERMINATION OF THE %E OF THE IRON SPECIES BY FAAS
In a centrifuge tube, 2.61 g of an aqueous solution containing only one of the iron species, Fe(II) (8.87 mg kg -1 ) or Fe(III) (7.94, 17.3, 39.7, or 86.7 mg kg -1 ), was added to 0.745 g of PEO1500, 0.642 g of Na 3 C 6 H 5 O 7 , and 3.60 mg of phen (5.00 mmol kg -1 of phen).This mixture was stirred, centrifuged, and left in a thermostatic bath at the abovementioned temperature, following which an ATPS with 2.00 g of each phase was obtained.Aliquots of the TP and BP were collected and adequately diluted, and their iron concentration was  measured by FAAS.The %E of the iron species in the TP and BP was calculated using the Eq. ( 1a) and Eq. ( 1b), respectively.

DETERMINATION OF IRON SPECIES IN WATER SAMPLES
Firstly, water samples doped with both iron species were prepared.The concentration of Fe(II) was always constant and equal to 8.87 mg kg -1 , while the Fe(III) concentration was varied as 7.94, 17.3, 39.7, or 86.7 mg kg -1 .Then, 0.642 g of the electrolyte (Na 3 C 6 H 5 O 7 ), 0.745 g of the polymer (PEO1500), 3.60 mg of the phen, and 2.61 g of the doped water sample were poured into a centrifuge tube.The contents of the tube were stirred and centrifuged; the tube was then left in a thermostatic bath as previously described.The mixture formed an ATPS with 2.00 g of TP and 2.00 g of BP.The TP and BP aliquots were then collected, appropriately diluted, and the total iron concentration in the top and bottom phases ( C T TP and C T BP ) were measured by FAAS.

STABILITY OF THE IRON SPECIES IN THE ATPS
A common problem encountered in the determination of the chemical species is the interconversion between the species.Especially in the case of iron, studies report that oxidation of Fe(II) to Fe(III) can occur in the presence of O 2 (Stumm and Lee 1961, King et al. 1995, Morgan and Lahav 2007, Jones et al. 2015).Hence, in this work, investigations were carried out to check whether Fe(II) got oxidized in the presence of dissolved oxygen in the ATPS, and if the ATPS could cause an interconversion of the species during the extraction procedure.Extraction studies of the species were carried out, with (by bubbling with nitrogen) and without the removal of oxygen dissolved in the ATPS (Butler et al. 1994).The TP extraction percentage for Fe(II), %E Fe(II)  TP , was obtained with the PEO1500+Na 3 C 6 H 5 O 7 +H 2 O ATPS, at pH of 6.00, and 10.0 mmol kg -1 of phen.The system containing Fe(II) remained in the thermostatic bath at 25 °C, and aliquots of the TP were taken at time intervals of 0, 0.25, 0.5, 2, 3, 4, and 20 h.The aliquots were adequately diluted and their metal concentration was determined by FAAS.In the presence or absence of O 2 , %E Fe(II) TP was found to be (90.6 ± 1.8)% or (91.7 ± 1.8)%, respectively; these values were similar for all investigated time intervals.These results indicate that Fe(II) did not undergo oxidation for at least 20 h.Therefore, it is not necessary to bubble the mixture with nitrogen; this makes the method easier by eliminating the nitrogen bubbling step from the procedure.

EXTRACTION BEHAVIOR OF THE IRON SPECIES IN THE ATPS
The complexation reaction between Fe(II) and 1,10-phenanthroline is well known (Vogel 1981, Marczenko et al. 2000).Studies report that the iron-phenanthroline complex (Fe(II)-phen) is formed by three molecules of the extractant and an iron atom (Lee et al. 1948).It is also recognized that the amount of extractant added to the ATPS influences the extraction behavior of the species (Patricio et al. 2011b, Da Cunha et al. 2016).Therefore, in this work, the effect of adding phen up to 100 times the molar quantity of iron, was investigated.
Figure 1 shows the %E of the iron species as a function of the phen quantity added to the PEO1500+Na 3 C 6 H 5 O 7 +H 2 O ATPS at pH of 6.00.In the absence of phen, Fe(II) and Fe(III) ions were practically not extracted for the TP (macromolecule-rich phase); in this situation, %E Fe(II) TP =(6.67 ± 0.60)% and %E Fe(III) TP (6.34±1.42)%.The presence of phen in the system did not influence %E Fe(III) TP ; however, %E Fe(II) TP gradually increased with increasing phen, reaching a maximum %E Fe(II) TP of (96.0 ± 3.9)% in the presence of 10.0 mmol kg -1 of phen.The low extraction efficiency of Fe(II) and Fe(III) in the absence of phen is related to the complexation of iron species with citrate anions that are mainly present in the bottom phase of the ATPS.At pH of 6.00 the citrate ions are deprotonated (pKas=3.128,4.761, and 6.396 (Harris 2013)); consequently, they are free to interact and form stable complexes with iron (Eq.( 2)), for which the formation constants are log K Fe(II)-Cit = 15.5 (Patnaik 2004) and log K Fe(III)-Cit = 25.0 (Patnaik 2004).Thus, the iron species stay primarily in the bottom phase (electrolyte-rich phase).
The addition of phen to the ATPS did not influence the extraction behavior of Fe(III), because Fe(III) interacts more strongly with citrate (present in large quantities in the BP) than with phen (predominantly present in the TP).This agrees with the Fe(III)-cit complex and Fe(III)-phen complex formation constant findings in literature (log K Fe(III)-Cit = 25.0 > log K Fe(III)-phen = 23.5)(Patnaik 2004).However, the addition of phen to the ATPS affected the extraction of Fe(II).Fe(II) preferentially reacts with the phen at the interface of the ATPS, because the Fe(II)-phen complex formation constant (log K Fe(II)-phen = 21.3)(Patnaik 2004) is greater than that of the Fe(II)-Cit complex (log K Fe(II)-Cit = 15.5)(Patnaik 2004).The Fe(II)-complex then gets transferred to the TP because of the interaction of the Fe(II)-phen complex with the pseudopolycations present in the TP.
In a pioneering study, da Silva and Loh (2000) proposed the existence of pseudopolycations to explain the formation of an ATPS.The authors determined the variation of the interaction energy between the ATPS-forming cations and the ethylene oxide segments of the polymer.They concluded that a biphasic system is generated because of the polymer-ion interaction.This interaction is governed by an increase in the enthalpy and entropy.
When phase separation occurs in the polymer-rich phase, cations are adsorved on the polymer chain, forming pseudopolycations.Hence, during the Fe(II) extraction process, the pseudopolycations attract the Fe(II)-phen complex because of the high negative charge density of the aromatic rings in phen, favoring the extraction of Fe(II) to the macromolecule-rich phase (TP).
With an increase in the quantity of phen in the ATPS, the %E Fe(II) TP also increases.This is because more complex molecules are formed and transferred to the TP by displacing the equilibrium toward the formation of more Fe(II)-phen complex.Thus, when 10.0 mmol kg -1 of phen is added, %E Fe(II) TP =(96.0±3.9)%.

INFLUENCE OF pH ON THE %E OF IRON SPECIES
Another factor that affects the formation of complexes between the iron species and phen, and consequently the preferential partitioning of the complex for one of the ATPS phases, is the pH medium.Hence, the partitioning of Fe(II) and Fe(III) species in the system was studied at pH values of 1.00, 6.00, and 12.0.
Figure 2 shows %E Fe(II) TP and %E Fe(III) TP as a function of the phen concentration in the PEO1500+Na 3 C 6 H 5 O 7 +H 2 O ATPS for different pH values.Fe(III), independent of the quantity of phen or the pH of the medium, was concentrated in the BP of the system, and reached %E Fe(III) TP equal to (6.64±2.56),(11.4±1.1), and (14.1±1.0)% on an average, for the pH values of 6.00, 1.00, and 12.0, respectively.At pH of 6.00, as aforementioned, the citrate ions are deprotonated and free to complex the Fe(III), concentrating Fe(III) specie in the BP.Therefore, a pH of 6.00 provided the lowest %E Fe(III)

TP
. At a pH of 1.0, the citrate ions are protonated, which increases the free-fraction Fe(III), and consequently the %E Fe(III) TP .At a pH of 12.0, %E Fe(III) TP is slightly higher, because part of the Fe(III) is complexed with hydroxyl-forming iron anionic aquohydroxy complexes ( Fe (OH) (3-x)   x ) that interact with the pseudopolycations present in the TP of the ATPS.
For all the investigated values of pH, %E Fe(II) TP increased with an increase in the quantity of phen, reaching a maximum %E Fe(II) TP of (96.0±3.9),(86.1 ± 0.6), and (85.9±0.6)% for pH values of 6.00, 1.00, and 12.0, respectively, when 10.0 mmol kg -1 of phen was added.The Fe(II) ions, at optimal phen concentrations, were quantitatively extracted for all investigated pH values.The highest %E Fe(II) TP of (96.0±3.9)% was obtained at pH of 6.00 in the presence of 10.0 mmol kg -1 of phen.In an acidic medium, protonation of the phen molecules occurs, which reduces the complexing efficiency of phen with Fe(II) and consequently, %E Fe(II)  TP .On the other hand, in basic media (pH = 12.0), the formation of Fe(II) aquohydroxy complexes ( Fe (OH) (2-x)   x ) decreases the Fe(II) available amount for complexation with phen.Hence, a medium with a pH of 6.00 was used as the optimal experimental condition for the development of the analytical method for iron speciation.This result agrees with literature data which affirm that the complex formation reaction between Fe(II) and phen is favored in slightly acidic solutions (Marczenko et al. 2000).

INFLUENCE OF THE NATURE OF THE ATPS-FORMING MACROMOLECULE ON THE %E OF THE IRON SPECIES
Studies have shown that the nature of the ATPS-forming macromolecule influences the extraction of metallic ions (Rodrigues et al. 2013, Patricio et al. 2016).In these works, the most commonly used macromolecules were PEO, L35, and L64.Normally, when a hydrophilic solute is partitioned, the systems containing PEO are more efficient in the extraction, while L35 and L64 are more efficient to extract the hydrophobic solutes.L35 or L64 are triblock copolymers, with a structure of (ethylene oxide) n -(propylene oxide) m -(ethylene oxide) n .Because of this structure, copolymer molecules aggregate above a specific temperature and concentration to form micelles.These micelles have a hydrophobic core (units of propylene oxide, PO) surrounded by a hydrophilic crown (units of ethylene oxide, EO) (Rodrigues et al. 2008, Patricio et al. 2016).Thus, an ATPS with a copolymer-rich phase will be more hydrophobic, while that with an polymer-rich phase will be more hydrophilic (da Silva et al. 2006, 2009, de Lemos et al. 2010, Virtuoso et al. 2012).
Figure 3 shows %E Fe(II) TP and %E Fe(III) TP as a function of phen concentration using the PEO1500 + Na 3 C 6 H 5 O 7 +H 2 O ATPS or the L64+Na 3 C 6 H 5 O 7 +H 2 O ATPS at pH of 6.00.In both systems, the phen amount added to the ATPS and the structure of the macromolecule had little effect on the Fe(III) extraction.The mean values of %E Fe(III) TP were (6.64±2.56)%and (14.1±1.0)% for the ATPS containing PEO and L64, respectively.This difference in the %E Fe(III) TP values is due to a difference in the concentration of Na 3 C 6 H 5 O 7 in the PEO ATPS (24.88% (m/m)) and L64 ATPS (13.54% (m/m)).The higher the concentration of Na 3 C 6 H 5 O 7 in the system formed by PEO, the more favorable is the formation of stable Fe(III)-citrate complexes.Thus, Fe(III) species concentrate in the BP (electrolyte-rich phase), resulting in a lower %E The extraction behavior of Fe(II) was dependent on the phen concentration and on the nature of the macromolecule present in the ATPS.There was a marked increase in %E Fe(II) TP with the increase of the phen amount added to the systems formed by both PEO1500 and L64.An increase in the amount of phen added to the system favors the formation of a higher amount of Fe(II)-phen complexes, which are subsequently transferred to TP (macromolecule-rich phase).However, the maximum %E Fe(II) TP of (96.0±3.9)%obtained for the ATPS containing PEO1500 was higher than the maximum %E Fe(II) TP of (51.6±3.0)%obtained in the system constituted by L64.The more efficient extraction in the presence of PEO1500 An Acad Bras Cienc (2018) 90 (2 Suppl. 1) is related to the structure of the macromolecule.PEO1500 is formed only by EO monomers, while L64 contains EO and PO monomers.Hence, the pseudopolycation amount formed in the L64 ATPS is lower than that formed in the PEO1500 ATPS, lowering %E Fe(II) TP in the L64 ATPS.

INFLUENCE OF THE NATURE OF THE ATPS-FORMING ELECTROLYTE ON THE %E OF THE IRON SPECIES
It is well established in the literature that the type of the ATPS-forming electrolyte influences the partitioning of solutes in the ATPS.Sulfate salts are widely used in metal extraction studies (Rodrigues et al. 2008, 2013, Patricio et al. 2011b), while the use of citrate salts is still rare (Da Cunha et al. 2016, Patricio et al. 2016).The effect of these anions on the separation of the iron species was also investigated.
Figure 4 shows %E Fe(II) TP and %E Fe(III) TP as a function of the phen concentration in the systems composed of PEO1500+Na 3 C 6 H 5 O 7 +H 2 O ATPSor PEO1500+Na 2 SO 4 +H 2 O ATPS at pH of 6.00.In the absence of phen, both iron species concentrated in the BP of the investigated systems.However, with the addition of phen, the Fe(II) species were preferentially partitioned toward the TP of the system, independent of the nature of the ATPS-forming electrolyte.A %E Fe(II) TP of (96.0±3.9)% was obtained for the ATPS composed of citrate, while a %E Fe(II) TP of (92.3±1.1)% was obtained for the ATPS composed of sulfate, where both systems were in the presence of 10.0 mmol kg -1 of phen.
The preferential partition of Fe(II) toward TP is due to the stronger interaction of Fe(II) and phen ( log K Fe(II)-phen = 21.3)(Patnaik 2004), as compared to the interactions of Fe(II) and citrate ( log K Fe(II)-Cit = 15.5 (Patnaik 2004) or Fe(II) and sulfate (log K Fe(II)-sulf.= 1.0) (Wells and Salam 1968).Hence, Fe(II) gets extracted to TP, because the Fe(II)-phen complex interacts more strongly with the macromolecules mainly present in TP due to the presence of pseudopolycations.When the PEO1500+Na 2 SO 4 +H 2 O ATPS was used, Fe(III) ions got extracted to TP, at a %E Fe(III) TP of (96.1±1.1)% in the presence of 10.0 mmol kg -1 of phen.Here, the interaction between Fe(III) and phen is more favored (log K Fe(III)-phen = 23.5 (Patnaik 2004) than the interaction of Fe(III) with the sulfate (log K Fe(III)-sulftate = 2.98 (Patnaik 2004) abundantly present in the BP.On the other hand, when the ATPS containing citrate was used, Fe(III) concentrated in the BP of the ATPS, and showed a %E Fe(III) TP of (6.64±2.56)%.This occurs because the interaction of Fe(III) with citrate (abundant in BP) is more intense than that of Fe(III) with phen.Hence, the PEO1500+Na 3 C 6 H 5 O 7 +H 2 O ATPS can be used for iron speciation.In this ATPS, Fe(II) is quantitatively extracted to TP, while Fe(III) is maintained in the BP.Note that it is not necessary to use 10.0 mmol kg -1 of phen to perform the speciation.When 5.00 mmol kg -1 of phen was added to the ATPS, a high %E Fe(II) TP of (95.1±1.0)%, and low %E Fe(III) TP of (7.68±0.50)%were obtained.Hence, the PEO1500 + Na 3 C 6 H 5 O 7 +H 2 O ATPS at a pH of 6.00 with a phen concentration equal to 5.00 mmol kg -1 was used as the optimal experimental condition to develop an analytical method for iron speciation.

SEPARATION FACTOR OF Fe(II) AND Fe(III) IONS
In a liquid-liquid extraction, it is possible to determine whether there is an effective separation between two species M and N, by determining the separation factor (S M, N ) (Park and Fray 2009).This factor is expressed in Eq. ( 3): where DM is the distribution coefficient for the species M, and DN is the distribution coefficient for the species N. The distribution coefficient for any species is expressed by Eq. ( 4):]where D M is the distribution coefficient for the species M, and D N is the distribution coefficient for the species N. The distribution coefficient for any species is expressed by Eq. ( 4): Figure 5 shows S Fe(II),Fe(III) as a function of the quantity of phen added to the PEO1500 + Na 3 C 6 H 5 O 7 + H 2 O ATPS at pH of 6.00.
In the presence of 10.0 mmol kg -1 of phen, the %E Fe(II) TP of (96.0±3.9)% and %E Fe(III) TP of (6.30±0.01)%resulted in a high S Fe(II),Fe(III) of 357.However, the use of 5.00 mmol kg -1 of phen also provides an expressive S Fe(II),Fe(III) of 233 with %E Fe(II) TP and %E Fe(III) TP equal to (95.1±1.0)% and (7.68±0.50)%,respectively.Iron speciation under this experimental condition provides economic and environmental benefits, due to a decrease in the quantity of phen used.In this work iron speciation in the presence of large quantities of Fe(III) was investigated.The extraction of Fe(II) was performed in a system formed by PEO1500+Na 3 C 6 H 5 O 7 +H 2 O, at a pH of 6.00 and with 5.00 mmol kg -1 of phen, in the presence of different concentrations of Fe(III) (0.101-1.01 mmol kg -1 ).The determination of Fe(II) in the ATPS phases was performed using the standard 1,10-phenanthroline method (Eaton and Franson 2005).Using the above methodology, %E Fe(II) TP and the extraction percentage of Fe(II) to BP (%E Fe(II) BP were obtained.It was observed that when the Fe(III) concentration in the medium increases from 0.101 to 1.01 mmol kg -1 , %E Fe(II) TP slightly varies from (98.4±0.7) to (101±1)%, while %E Fe(II) BP slightly varies from (1.65±0.67)% to (0.369 ± 0.183)%.These results demonstrate that it is possible to determine Fe(II) in the presence of Fe(III), since in these conditions, the phen reacts only with Fe(II).

DETERMINATION OF IRON SPECIES IN WATER SAMPLES
Firstly, only Fe(II) or only Fe(III) was added to some water samples.The ATPS formed by PEO1500+Na 3 C 6 H 5 O 7 +H 2 O at pH of 6.00 containing 5.00 mmol kg -1 of phen was used to separate the iron species in the water samples before the determination of the iron species by the non-selective and low-cost FAAS technique.The percentage extraction of each specie in each ATPS phase was then determined using Eq.(1a) and Eq. ( 1b).The experiments were repeated for many days and at different concentrations of the iron species to evaluate the variation in the extraction percentages (Table II).In addition, known amounts of Fe(II) and Fe(II) were added to the different water samples, and recovery was determined (Table III).The known concentration of Fe(II) was called C S,ad Fe(II) , while that of Fe(III) was denoted as C S,ad Fe(III) .
The results of Table II show  (5) The terms C T TP and C T BP are divided by 2 because the mass of ATPS is 4.00 g, but the masses of TP and BP are 2.00 g each.
where m S is the mass of the water sample added in the ATPS, and m ATPS is the mass of the ATPS.
The known concentration values of Fe(II) and Fe(III) ( C S,ad Fe(II) and C S,ad Fe(III) ) were very close to the concentration values obtained ( C S,ob Fe(II) and C S,ob Fe(III) ) by the proposed method, reaching recoveries ranging from 90% to 106%.These results demonstrate the efficacy of an ATPS for iron speciation and the possibility of using FAAS for the determination of species determination.
Some researchers used UV-Vis spectrophotometry to determine the Fe(II) and Fe(III) species (Jamali et al. 2016, Chen et al. 2015, Suárez et al. 2015, Bazmandegan et al. 2015, Silva et al. 2018), since it is one of the cheapest analytical technique (Silva et al. 2018, Pena-Pereira et al. 2011).However, in the present work was used FAAS because it is necessary less steps to prepare solutions and analytical curve, reducing reagents consume and time spent.In addition, it has other attractive green and useful characteristics as summarized in Table IV.As can be seen, the proposed method has several advantages comparing to other reported methods for speciation of iron species.Unlike the others, it does not require organic solvents, corrosive, reducing or oxidizing agents.In addition, it can be performed without the removal of dissolved oxygen in the ATPS and demanding a short stirring time of 180 s.f Fe(III) concentration obtained by the proposed method.g Recovery ± standard deviation, n = 5.

CONCLUSIONS
An economically viable and environmentally friendly method was proposed for the separation and determination of iron species.This method uses aqueous two-phase systems for the separation of the species, which provides several advantages like the use of non-toxic, biodegradable and recyclable components, ease of operation, low-cost, and rapid phase separation without the formation of stable emulsions.In addition, the non-selective FAAS technique can be used for chemical speciation.
The studies carried out in this work showed that, through a selection of the ATPS-forming macromolecule and electrolyte, and a control of the quantity of phen and the pH of the medium, it is possible to separate Fe(II) and Fe(III) without interconversion of the species during the separation and determination.The ATPS formed by the PEO1500 macromolecule and the Na 3 C 6 H 5 O 7 electrolyte was found to be most efficient for the iron speciation.For extractant addition, in the absence of phen, the ions concentrated in the bottom phase, giving %E values lower than 15.0%.By the addition of phen, Fe(II) was extracted to the top phase, while Fe(III) was maintained in the bottom phase.
The obtained results can be considered to be extremely satisfactory, because ATPS-FAAS coupling was employed for the first time in chemical speciation.This approach can be considered more environmentally friendly than those resorting to organic solvents and at the same An Acad Bras Cienc (2018) 90 (2 Suppl. 1)  time, expands it realm of application for speciation of other elements.Furthermore, the synergy arising from the combination of the beneficial characteristics of flame atomic absorption spectrometry, such as attractive cost, simple operation, wide availability in many laboratories, with the advantages of an aqueous two-phase extraction system makes this work an attractive and promising starting point for future studies and developments in chemical speciation.

Firstly
that the varying the concentration of Fe(III) in the ATPS does not affect the extraction percentage of the species.It can be concluded that when an ATPS is used in the extraction, a %E Fe(II) TP of 118%, %E Fe(II) BP of 8.10, %E Fe(III) BP of 120%, and %E Fe(III) TP =14.4% do not vary.Hence, these percentage extraction values were used to determine the concentration of each iron specie in the real samples.were substituted in the Eq.(5) and Eq.(6).Secondly, the total iron concentration was determined in the TP ( C T TP ) and in the BP ( C T BP of the ATPS by FAAS.Then, the obtained values of C T TP and C T BP were substituted in Eq. (5) and Eq.(6).It was then possible to obtain the concentrations of Fe(II) (C ATPS Fe(II) ) and Fe(III) (C ATPS Fe(III) ) in the ATPS with the water samples.Finally, the values of C ATPS Fe(II) and C ATPS Fe(III) were substituted in Eq. (7)and Eq. (8) to obtain the concentration of Fe(II) ( C S,ob Fe(II) ) and Fe(III) (C S,ob Fe(III) ) in the water sample.

TABLE I Concentration of components in the ATPS.
M: macromolecule, E: electrolyte.

TABLE II Percentage extraction of Fe(II) and Fe(III) using PEO1500+Na 3 C 6 H 5 O 7 +H 2 O ATPS at a pH of 6.00 with 5.00 mmol kg -1 of phen.
Known concentration of Fe(II) in doped samples.b Known concentration of Fe(III) in doped samples.c Mean extraction percentage of Fe(II) to TP ± standard deviation, n = 5. d Mean extraction percentage of Fe(III) to TP ± standard deviation, n = 5. e Mean extraction percentage of Fe(II) to BP ± standard deviation, n = 5. f ean extraction percentage of Fe(III) to BP ± standard deviation, n = 5. a

TABLE III Determination of Fe(II) and Fe(III) concentration in water samples.
a Known concentration of Fe(II) in doped samples.b Known concentration of Fe(III) in doped samples.c Total iron concentration determined in the TP.d Total iron concentration determined in the BP. e Fe(II) concentration obtained by the proposed method.