Partitioning of perrhenate anion by aqueous two-phase systems using design of experiments methodology

Accepted Manuscript Partitioning of perrhenate anion by aqueous two-phase systems using design of experiments methodology

Lucy Muruchi, Yecid P. Jimenez PII: DOI: Reference:

S0167-7322(17)32232-8 doi:10.1016/j.molliq.2017.10.091 MOLLIQ 8053

To appear in:

Journal of Molecular Liquids

Received date: Revised date: Accepted date:

22 May 2017 17 October 2017 19 October 2017

Please cite this article as: Lucy Muruchi, Yecid P. Jimenez , Partitioning of perrhenate anion by aqueous two-phase systems using design of experiments methodology. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Molliq(2017), doi:10.1016/j.molliq.2017.10.091

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Partitioning of perrhenate anion by aqueous two-phase systems using design of experiments methodology

Lucy Muruchi, Yecid P. Jimenez* Department of Chemical and Mineral Process Engineering, Universidad de Antofagasta, Av. Angamos 601,

IP

T

Antofagasta, Chile

CR

Abstract

Rhenium is a valuable metal in very low concentration and it is being increasingly used in industry due to its

US

multiples properties, mainly in the aeronautics and oil industries. For this work, the PEG/CuSO4 aqueous two phase system (ATPS) was used for the partition of perrhenate anion. For the partition tests the NH4ReO4

AN

salt was used. The effects of temperature, pH, salt concentration, PEG concentration and their interactions on

M

the partition were analyzed by using experimental design 2 k. Moreover, density, kinematic viscosity and sound velocity were determined for the top and bottom phases. The optimal conditions obtained by the

ED

model for the perrhenate partition in the ATPS formed by PEG and CuSO 4 were 13.7% PEG 4000 (w/w), 13%

PT

CuSO4 (w/w) and 35 °C, obtaining a distribution coefficient, KReO4- , of 7.72 ± 0.31. Also, an increase of KReO4with increasing PEG and CuSO4 concentration was observed, i.e. their effects are significant on the

CE

perrhenate partition. On the other hand, the effects of temperature and pH do not have high significance on

AC

the partition. Of all the studied parameter, the pH had less significance on the partition. Thus, to simplify the statistical model the pH was not considered in the model.

For both phases, only density is inversely proportional to pH. With exception of the sound velocity of bottom phase all properties for both phases are inversely proportional to temperature. Moreover, all properties are proportional with the PEG and CuSO4 concentrations in both phases.

1

ACCEPTED MANUSCRIPT The partitioning behavior was evaluated by an experimental design, applying statistical regression analysis and analysis of variance (ANOVA). The statistical model for distribution coefficient as function of the parameters with significant effects was obtained. A good agreement between experimental and correlated data was observed (R-Squared=0.930).

IP

T

Key words: Aqueous two-phase system; partition; PEG 4000; perrhenate; experimental design.

AC

CE

PT

ED

M

AN

US

CR

* [email protected]

2

ACCEPTED MANUSCRIPT 1. Introduction Rhenium is a valuable and rare metal present in the earth crust at very low concentration (0.4 mg/t) [1]. In last years, its use in industry, especially in high performance alloys for aerospace application have increased. Metallic rhenium also can be alloyed with other metals to form acid- and heat-resistant materials, and it is

T

also used for electrical and electronic filaments, electric contacts, heaters and thermocouples, due to its high

IP

melting temperature (3459 K). Rhenium compounds are used in electronic and semiconductor

CR

manufacturing and can be employed as a protective coating on other metals [2].

US

Due to its multiple valences (− 1 to + 7), can be used for catalys reactions in the petrolium industry as rhenium metallic or salt. Catalyst reactions containing rhenium compounds are usually prepared from

AN

ammonium perrhenate (NH4ReO4) and and sodium perrhenate (NaReO4). Some applications of this kind of reactions are: production of gasoline of high octane, rhenium sulfides obtained from ammonium perrhenate,

M

to dehydrogenation of ethyl alcohol and higher alcohols to aldehydes and ketones.

ED

Rhenium is recovered from molybdenite concentrates, which are obtained as a by-product of copper

PT

refining. These molybdenite concentrates are roasted to obtain the molybdenium oxide. During the roast process, rhenium is volatized as heptoxide, then, it is scrubbed or leached from the flue gases and dust for its

CE

conversion to perrhenic acid. This acid solution is purified by solvent extraction or ionic exchange. Finally,

AC

ammonium perrhenate is obtained by crystallization. From this ammonium salt, other compounds can be produced, e.g. rhenium metallic by reduction with hydrogen, rhenium heptoxide by evaporation of the salt on a carrier and subsequent calcinations [3]. Also, ammonium perrhenate (NH 4ReO4) is obtained from the leaching solutions of the copper ore, molybdenum and lead production, and from the recycle solutions of rhenium-containing residues [1, 4]. In these cases, the solvent extraction of pregnant leaching solution (PLS) is an effective method to separate selectively different metal ions. However, according to Rogers et al. [4] even with today's environmental standards, there are still a number of extraction systems that utilize toxic

3

ACCEPTED MANUSCRIPT and flammable organic diluents. Also, the cost of the solvent system can become very expensive if the diluent is used with a highly selective extractant, not to mention the costs of safely designing a system to operate with a diluents of these characteristics [5] .

Aqueous two-phase system (ATPS) compared with the traditional organic solvent extraction process is a

T

new green method, which has the advantage of being non- toxic, nonflammable and low-cost [6]. The

IP

possibilities of linear scale up, ease of use and rapid phase separation without the formation of stable

CR

emulsions are other advantages of the ATPSs. The APTS have been widely used in extracting and separating

US

of bioactive substance [7-9], organic small molecule and metal cations. However, only a few results have been reported in the extraction and separation of metal anion by ATPS.

AN

Nowadays, the optimization of partitioning processes in PEG–salt ATPS is widely carried out by a

M

multivariate statistical technique called design of experiments (DoE) which is based on a statistical factorial experimental design concept [10], consisting of the performance of a few experiments at a particular factor

ED

level combination. The performance of a designed experiment is usually related to the determination of the

PT

effects of changed controllable input factors on the corresponding varied output responses.

CE

The partition of ions in ATPS is influenced by pH, temperature, surface properties, concentrations and the types of polymers and salts employed. In the present study, the phase-forming salt, CuSO4, and the

AC

poly(ethylene glycol) (PEG) with molar mass of 4000 were selected. Since, as mentioned above, one of the major sources to produce NH4ReO4 are the PLSs in sulfate medium of porphyry copper ores [11], [12]. Moreover, as a continuation of previous works, where the PEG 4000 was used, it is interesting to study the probability of partitioning perrhenate anion in an ATPS formed by poly(ethylene glycol) and copper sulfate that could constitute an alternative process for the extraction of perrhenate anion [13], [14], [15]. Scott K. et al [16] studied the partition of ReO4- on ATPSs composed of PEG 2000-H2O and as forming phase salts: Na2MoO4, Na2WO4 and NaOH, at 25°C. The authors concluded that increasing the forming

4

ACCEPTED MANUSCRIPT phase salt concentrations, the distribution coefficient increases. The best values for distribution coefficient were reported for the ATPS formed with NaOH. Also, the authors have simulated conditions of alkaline solutions with 6 M NaOH and 0.1 M Na2MoO4 or 0.1 M Na2WO4 for the partition of ReO4-, concluding that these conditions not presents problems in the partition until 0.1M of NaReO 4. At the end, the distribution

T

coefficient seems to be controlled by NaOH with small contribution of additional phase forming salts

IP

(Na2MoO4 or Na2WO4). However, the authors not present the type of study proposed here, since, the present

CR

work studies the effect of the temperature, pH, salt and PEG concentrations in the partitioning of the ReO 4anion. . Also, in this work, different mass molar for PEG (4000) is used and the phase forming salt is copper

US

sulfate. Other important difference is the application of the experimental design methodology. Since, the

AN

influence of process parameters such as PEG and salt concentrations, temperature and pH of the system on

M

the partitioning of ReO4- anion was studied by 2 k factorial design.

ED

2. Experimental

PT

2.1 Chemicals

Synthesis-grade polyethylene glycol with an average molar mass of 4000 (3500-4500) g∙mol-1 and for the

CE

ReO4- anions, analytical grade reagent ammonium perrhenate (NH4ReO4) with a purity 99.98% mass fraction

AC

were procured from Sigma-Aldrich and Molymet, respectively. The water content in the PEG 4000 was determined by the Karl Fisher method and was 0.0031 (mass fraction). This water content was neglected of the total solution mass. Sulfuric acid, with a purity of 0.95-0.97 mass fraction (density: 1.84 g∙cm-3) and Copper Sulfate pentahydrate (CuSO4*5H2O) for analysis were procured from Merck. All chemicals were used without further purification. Milli- Q quality deionized water with electrical conductivity 0.054 μS∙cm-1 was used in all experiments.

5

ACCEPTED MANUSCRIPT 2.2 Apparatus and procedure The liquid-liquid equilibria data of the ATPS formed by CuSO4+H2O+PEG 4000 and reported by Claros et al. [15] was used to the partition tests. The partitioning of perrhanate anion was carried out by using a 2 k experimental design, where “k” is the number of factors or parameters and for this work these considered

T

factors were temperature, pH, salt and PEG concentrations of the ATPS. The low and high levels for each

IP

factor in the experimental design are presented in the Table 1. The chosen range for the parameters of

CR

temperature, salt and PEG concentration was based on the phase diagram. In the case of pH, preliminary

US

tests showed a precipitate in the system at pH values higher than or equal to 3 .

The partitioning of ReO4- in a PEG/CuSO4 ATPS was carried out by using a full factorial design (FFD) 2 k

AN

involving four variables: temperature, pH, salt and PEG concentrations of the system at two levels. Table 1 shows the design matrix of the variables and the most relevant results, including the partition coefficient of

M

ReO4-, KReO4-, densities and sound velocities for both of phases. Design expert software 7.0 to design the

ED

experiments, obtain the statistical model and analyze the results was used. The factor combination resulted

AC

CE

PT

in 32 experiences (including repetitions).

6

ACCEPTED MANUSCRIPT Table 1. Low and high levels for each factor in the experimental design provided by software, pH, temperature, T, CuSO4 mass fraction, w1, PEG 4000 mass fraction, w2, and response variables, distribution coefficient for ReO4- anion (K ReO4-), bottom phase density, ρbp, upper phase density ρup, sound velocity of bottom phase, sbp and sound velocity of upper phase sup.

Std Run

9 3 15 11 1 7 19 27 23 17 21 13 25 5 29 31 30 26 18 20 24 14 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

T P

Factor 1 Factor 2 Factor 3 Factor 4 Response 1 Response 2 Response 3 Response 4 Response 5

Block

Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1 Block 1

A:T/°C

B: pH

C: w1

D: w2

K ReO4-

ρbp/g·cm-3

35 sd 45 45 45 35 45 45 45 45 35 35 35 35 35 35 45 35 35 35 45 45 35 35

1 1 2 1 1 2 1 1 2 1 2 2 1 2 2 2 2 1 1 1 2 2 1

0.13 0.112 0.13 0.13 0.112 0.112 0.112 0.13 0.112 0.112 0.112 0.13 0.13 0.112 0.13 0.13 0.13 0.13 0.112 0.112 0.112 0.13 0.13

0.108 0.108 0.108 0.108 0.108 0.108 0.137 0.137 0.137 0.137 0.137 0.108 0.137 0.108 0.137 0.137 0.137 0.137 0.137 0.137 0.137 0.108 0.108

7.06 4.42 5.9 6.15 4.25 4.56 5.06 6.65 5.43 6.07 5.77 6.64 7.6 4.66 7.73 6.72 7.85 7.74 5.91 4.83 5.84 6.21 6.52

1.20860 1.17414 1.19730 (1.19940 g·cm-3) 1.18272 1.17124 1.19217 1.21942 1.20860 1.18959

C A

E C

T P

D E

M

7

C S

U N

A

I R

ρup/g·cm-3

1.20274 1.20023 1.20628 1.22980 1.18016 1.22783 1.21770 1.22968 1.22976 1.20258 1.19165 1.19022 1.20569 1.20896

1.10332 (g·cm-3) 1.09233 1.08960 1.09620 1.10458 1.08479 1.09364 1.09896 1.08503 1.09972 1.09017 1.09476 1.10335 1.09105 1.09362 1.08492 1.10251 1.10442 1.09847 1.09162 1.08462 1.09325 1.10276

sbp/m·s-1

sup/m·s-1

1607.39 (m·s-1) 1599.59 1613.17 1611.77 1593.67 1600.65 1608.80 1621.81 1611.15 1605.38 1607.34 1609.65 1620.09 1595.96 1622.82 1624.12 1620.62 1619.87 1605.55 1608.79 1610.44 1609.68 1607.68

1664.29 (m·s-1) 1647.57 1662.18 1659.82 1646.82 1650.83 1657.26 1668.26 1661.07 1663.23 1667.37 1667.72 1676.97 1652.97 1679.36 1669.98 1677.96 1676.34 1663.62 1659.23 1661.24 1666.36 1664.04

ACCEPTED MANUSCRIPT 16 24 Block 1 28 25 Block 1 32 26 Block 1 22 27 Block 1 4 28 Block 1 12 29 Block 1 6 30 Block 1 8 31 Block 1 2 32 Block 1 The standard uncertainties u, are

45 2 0.13 0.108 6.02 45 1 0.13 0.137 7.44 45 2 0.13 0.137 7.34 35 2 0.112 0.137 6.21 45 1 0.112 0.108 5.14 45 1 0.13 0.108 6.03 35 2 0.112 0.108 4.89 45 2 0.112 0.108 4.51 35 1 0.112 0.108 4.58 u(T) = 0.01 K for the density and sound velocity,

1.19774 1.08929 1613.23 1661.17 1.21918 1.09511 1622.59 1668.99 1.21716 1.08914 1624.27 1670.07 1.20061 1.09028 1607.64 1667.29 1.17302 1.09055 1599.84 1648.63 1.19972 1.09672 1611.42 1659.79 1.18018 1.09059 1596.35 1653.67 1.17736 1.08579 1603.80 1653.39 1.18103 1.09845 1594.37 1649.72 u(T) = 0.01 K for coefficient distribution, u(ρbp) = 8·10-5g·cm-3,

T P

u(ρup) = 9·10-5 g·cm-3, u(sbp) = 0.06 m·s-1, u(sup) = 0.08 m·s-1, and u(K ReO4-) = 0.3.

C S

U N

A

D E

M

T P

E C

C A

8

I R

ACCEPTED MANUSCRIPT The liquid-liquid equilibria data of ATPS CuSO4+H2O+PEG 4000 reported by Claros et al. [14] was used for the partitioning tests. With increasing temperature, the binodal curves move towards the origin, for this reason the selected concentrations at 298.15 K were also suitable for the other temperatures of partition. Typically, 35 g of a solution of known composition of PEG, CuSO 4, water (ATPS forming components) and a

T

NH4ReO4 constant amount (0.4g) to partition was prepared in jacketed vessels by mass using a Mettler

IP

Toledo analytical balance (model AX204) with a precision of ± 0.1 mg. Then, these solutions were agitated

CR

until complete dissolution using a VWR® Professional magnetic stirrer with operating conditions from 278.15 to 313.15 K and a speed range from 60 to 1600 rpm. A thermostatically controlled Julabo bath F25-ME

US

Refrigerated/Heating Circulator was used to maintain a constant temperature, with working temperature

AN

range 245.15-473.15 K and an uncertainty of ± 0.01 K. After the complete dissolution, the pH system was adjusted by adding sulfuric acid; the pH values were measured using an Accumet pH meter 50 with a

M

measurement range from -2 to 20 between 268.15 and 378.15 K and a relative accuracy of ± 0.002. The

ED

obtained biphasic solutions were allowed to settle for 20 minutes at constant temperature. Longer periods of agitation and sedimentation did not affect the results.

PT

Once equilibrium was reached, the top and bottom phases were withdrawn using syringes and needles to

CE

determine the rhenium concentration of each phase and measure the physical properties. Densities and sound velocities of the solutions were measured using a density and sound velocity meter (Anton Paar DSA

AC

5000M, measuring range from 0 to 3 g·cm -3, resolution 1·10-6 g·cm-3) with a precision of ± 0.001 K. The obtained uncertainties were ±1·10-5 g·cm-3 and ±0.03 m·s-1 for the density and sound velocity, respectively. The density and sound velocity meter was calibrated using air and deionized water as reference substances. Three milliliters of solution for each measurement were used, and all measurements were conducted in triplicate. The reported data represent the average of these measurements. The kinematic viscosities were measured with calibrated micro-Ostwald viscometer. A Schott-Gerate automatic measuring unit (model AVS 310) equipped with a thermostat (Schott-Gerate, model CT 52) in

9

ACCEPTED MANUSCRIPT which the temperature was regulated to within ± 0.05 K was used for the measurements. The uncertainty of the kinematic viscosity measurements was ±5·10-3 mm2·s-1. The measurements were also conducted in triplicate, and 2 ml were used for each experiment.

Rhenium concentration was determined by atomic absorption spectroscopy (AAS). The AAS measurements

IP

T

were performed using a Varian Atomic Absorption Spectrophotometer, model 220. The NH4+ was

CR

determined by ion selective method.

Where

and

(1)

AN

US

Results were evaluated by the distribution coefficient in the aqueous two-phase systems, defined as:

are concentrations of

M

3. Results and discussion

in top and bottom phase, respectively.

ED

3.1 Effects of independent variables on the partition of ReO 4- anion

PT

In biotechnology area, the five independent variables most commonly used are the molecular weight of PEG, PEG concentration (wt. %), phase-forming salt concentration such as phosphate or citrate (wt. %), pH and

CE

additional neutral salt like NaCl or KCl (wt. % or M) [10], while the partition coefficient (K), product recovery/activity yield (Y, %) and purification factor (PF) are mostly set as the response variables (dependent

AC

response variables). Furthermore, other independent variables, such as temperature (°C) and biomolecule/ bioligand load/concentration (mg/mL or wt. %) and response variables, like the selectivity (S), are also considered. In this work, the parameters of PEG concentration, phase forming salt concentration, temperature and pH system were considered. The effect of these variables on the behavior of distribution coefficient is presented in the following sections.

10

ACCEPTED MANUSCRIPT 3.2 Effect of system pH According to the data analysis, it was evident that the pH had no significant effect on the distribution coefficient. Thus, for simplify the statistical model, the pH is not considered in the model. However, although statistically the pH effect is not significant compared with the other factors, there is a level change

IP

T

indicating proportionality between KReO4- and pH. Figure 1 shows this behavior for the experimental results.

6.50 5.40

CR

5.00

5.50

4.80

pH 1

4.60

K

K

6.00

w1=0.112 w2=0.108

5.20

5.00

pH 2

4.50

4.20

pH 2

w1=0.112 w2=0.137

US

4.40

pH 1

4.00

4.00 35

40

30

45

T/°C

M

a) Effect of pH at low PEG concentration, w2

AN

30

35

40

45

T/°C

b) Effect of pH at high PEG concentration, w2

ED

Figure 1. Effect of pH at low and high PEG concentration, w2, on KReO4-, at different temperatures and low

PT

level of CuSO4 concentration (w1 = 0.112) suggested for actual experiences.

CE

Similar behavior was reported by Yongqiang et al. [6], who studied the extraction of molybdate anion by ATPS. The authors concluded that the pH values define the specie present in the system, finding that

AC

polymolybdate anion with more relative high charge density existing at pH 2-4, so the extraction rates and distribution coefficients of molybdenum are higher and proportional with the pH.

Also, many authors deduce that the two phases of an ATPS have a defined net charge [17], [18], [19], [20], [21]. Oliveira et al. [22] and Rodrigues et al. [23] reported that the ATPS formed by a polymer, salt and water has an electrostatic differential potential between the two phases due to salt. The top phase due to the formation of polications has a positive charge, these polications are formed by the solvation of the polymeric macromolecule (e.g. PEG) with the cations of the salt-forming phase. The top phase may have a positive 11

ACCEPTED MANUSCRIPT charge of high or low intensity depending on the cation. The cation and the polymer interact, releasing some water molecules that were solvating them, in a process driven by the entropy increase. This cation binding continues as more electrolyte is added, until a saturation point, after what no more entropy gain may be achieved and the phase splitting becomes more favorable. After this saturation point, the addition of more

T

salt would lead to a higher concentration in the bulk than around the polymer; this may explain the salting-

IP

out effects [24]. Banik et al. [25] reported that the partition of molecules vary with the charged of the phase-

CR

forming polymer derivative to the ATPS. The main principle in this is the concept that like repels like and unlike attract each other. Thus, when the molecule has an opposite charge to that of the polymer, the K value

US

increases substantially and vice versa. In this way, electrostatic interactions between charged molecules and

AN

the phases contribute to the partitioning process. Silva et al. [21] studied the partition of glucose-6-phosphate dehydrogenase (G6PDH) and hexokinase (HK) in ATPSs formed by PEG and NaH 2PO4 or K2HPO4. The

M

authors reported that the top phase rich in PEG has a positive charge and the bottom phase rich in salt has a

ED

negative charge, and concluded that the partition of biomolecules depends on their net surface charge. For both enzymes, with increasing the pH, the partition coefficient increases, so, it was concluded that by

PT

modifying the pH of the system the net surface charge also change. According to Asenjo [26], the pH may

CE

alter the chemical groups located in the lateral chain of protein amino acids, which produces a modification of the net charge of the biomolecules, changing their solubility. Also, the partition of the molecules depends

AC

on their isoelectric points (IPs). The IP of a molecule is defined as the pH at which it has the same positive and negative charge numbers. Naganagouda et al. [19], Rosa et al. [20] and Zhang et al. [27] reported that the molecule surface has a net negative charge at pH values above IP, thus in the partition process of molecules, these are partitioned to the top phase. This situation is reversed when the pH value is lower than the IP.

In this work, ReO4- is a small anion with high charge density and with an isoelectric point below pH 1 that would explain the slightly high distribution coefficients at pH 2.

12

ACCEPTED MANUSCRIPT For the experimental results, Figure 1 shows the interaction between pH and PEG concentration. The pH effect is smaller at the low level of the PEG concentration (w 2 = 0.108) than at the high level of the PEG concentration (w2 = 0.137), obtaining better result at high PEG concentration (w 2 = 0.137) and pH 2.

3.3 Effect of temperature on the system

T

Figure 2 shows that for the lower PEG concentration, the effect of temperature, according to the slopes, is

K Partition coef f icient B- 1.000 B+ 2.000

B: pH

US

7.90

X1 = A: Temperature °C X2 = B: pH

Actual Factors C: CuSO4, w/w = 0.12 D: PEG 4000, w/w = 0.14

6.05

5.13

35.00

37.50

40.00

42.50

6.05

5.13

4.20

37.50

40.00

42.50

45.00

A: Temperature °C

b) Effect of temperature on KReO4- at high PEG

PT

a) Effect of temperature on KReO4- at low PEG

concentration

CE

concentration

6.98

35.00

45.00

ED

A: Temperature °C

M

4.20

K Partition coefficient

6.98

AN

Actual Factors C: CuSO4, w/w = 0.12 D: PEG 4000, w/w = 0.11

K Partition coefficient

B- 1.000 B+ 2.000 X1 = A: Temperature °C X2 = B: pH

Interaction

Design-Expert® Sof tware

B: pH

7.90

CR

Interaction

Design-Expert® Sof tware K Partition coef f icient

IP

less than at high level of PEG concentrations.

AC

Figure 2. Effect of temperature at low and high PEG concentration on K ReO4- at low level of CuSO4 concentration (w1 = 0.112) suggested by the model.

The line slopes in Figure 2 shows that in all conditions at low temperatures, better KReO4- are obtained.

The inversely proportional relation between coefficient distribution and temperature is reported by other authors. Yongqiang et al. studied the selective extraction and separation of molybdenum (VI) from aqueous media by APTS formed by Triton X-100 + (NH4)2SO4 + H2O [6]. At the beginning, the distribution coefficient with increasing temperature shows an increase and then a decrease. The authors concluded that

13

ACCEPTED MANUSCRIPT the distribution coefficient decreases due to higher water content and larger phase volume [28], leading to lower molybdenum concentration in TX-100-rich micellar phase. However, after 303.15 K, non-polar conformers are strengthened and begin to dominate the properties of polymers based on ethylene oxide units. Polymolybdate anion is polarity; increasing incompatibility leads to decrease of the distribution

T

coefficient with increasing temperature, indicating that the polymolybdate anion transferring from the salt-

IP

rich phase to the TX-100-rich micellar phase is an exothermic process. On the contrary, the phase-separation

CR

process is endothermic [29].

US

Kammoun et al. [16] found similar behavior for the partition of CBS 819.72 α-amylase by ATPSs. The authors concluded that in ATPSs conformed by PEG, buffer citrate/phosphate and additives like NaCl or KCl, the

AN

partition is better at 4°C than at 20°C [17].

M

However, for the biomolecules partitioning, some authors reported that the partition coefficient increases monotonically with increasing temperature [19], [18]. Moreover, the authors reported that the binodal curve

ED

gradually moves towards the origin as the temperature increases [29]. This indicates that as temperature

PT

increases the PEG concentration increases in the top phase and the phosphate salt concentration in the bottom phase decreases. The preferential attraction of the water molecules to the polar salt surface instead of

CE

the PEG as temperature increases reduces the solubility of PEG in the bottom phase. This is known as the

AC

“salting-out” effect, which becomes stronger with increasing temperature. A similar conclusion can be made for the protein molecules because they are also less polar than salt. For this reason, the affinity of the biomolecule by the top phase increases.

The most important process variables such as phase-forming polymers or salts, pH, buffer, ion strength and temperature depend on the characteristics of the target protein to be partitioned such as hydrophobicity, molecular size, electrochemical properties and molecular conformations [30], [19], for these reasons some distribution coefficient are or not proportional with the temperature. 14

ACCEPTED MANUSCRIPT Due to the above, and according to the distribution coefficient obtained, the polarity of the anion ReO4makes that its solubility in the bottom phase increases with increasing temperature. At the same time, the top phase becomes more hydrophobic. Thus, the anion is partitioned to the bottom phase with increasing temperature; this would explain the decrease of the distribution coefficient.

T

Figure 2 also shows that the effect of PEG concentration is very significant on the distribution coefficient

Interaction

Design-Expert® Sof tware

K Partition coef f icient

B- 1.000 B+ 2.000

B: pH

7.90

US

6.98

X1 = A: Temperature °C X2 = B: pH

Actual Factors C: CuSO4, w/w = 0.13 D: PEG 4000, w/w = 0.12

6.05

AN

Actual Factors C: CuSO4, w/w = 0.11 D: PEG 4000, w/w = 0.12

K Partition coefficient

B- 1.000 B+ 2.000

X1 = A: Temperature °C X2 = B: pH

Interaction

Design-Expert® Sof tware

B: pH

7.90

5.13

35.00

37.50

40.00

42.50

45.00

6.05

5.13

35.00

37.50

40.00

42.50

45.00

A: Temperature °C

ED

A: Temperature °C

6.98

4.20

M

4.20

K Partition coefficient

K Partition coef f icient

CR

IP

Figure 3 shows a similar behavior for the effect of the salt concentration on the distribution coefficient.

b) Effect of high salt concentration on KReO4-

PT

a) Effect of low salt concentration on KReO4-

w2 = 0.12 of PEG suggested by the model.

AC

CE

Figure 3. Effect of salt concentration (Cu2SO4) on distribution coefficient at low and high temperature, and

The proportionality between salt phase-forming and distribution coefficient is attributed to the salting-out effect. In this sense, for the selective extraction and separation of molybdenum (VI) from aqueous media by APTS formed by Triton X-100 + (NH4)2SO4 + H2O, Yongqiang et al. [5] reported that the distribution coefficient

of

molybdenum

initially

increases

and

then

remains

invariant

with

increasing

(NH4)2SO4 concentration, due that the salt is kosmotropic and has a large free energy of hydration. This salt has a strong salting-out ability and becomes stronger with increasing concentration, which results in the decrease of free water concentration, making more the chaotropic polymolybdate anion dehydrate, and 15

ACCEPTED MANUSCRIPT increasing its hydrophobic nature and enhancing its extracting performance from salt-rich phase to the TX100-rich micellar phase. Also, this produces low water content and small phase volume in TX-100 micellar phase. Wu et al. [30] reported similar behavior for the extraction and separation of butyric acid from clostridium tyrobutyricum fermentation broth in PEG/ Na2SO4 aqueous two-phase system [31]. The authors

T

also concluded that with an additional increase of Na2SO4, the volume and the water content of the sodium

IP

sulfate-rich phase increases, driving PEG molecules bound with butyric acid , from the PEG-rich phase to the

CR

sodium sulfate-rich on, resulting in a drop in butyric acid yield.

US

According to the above considerations, CuSO4 is a kosmotropic salt with a large free energy of hydration, whose values are: ΔG°hyd(Cu2+) = −2010 kJ·mol−1 and ΔG°hyd(SO42−) = −1080 kJ·mol−1 at 298.15 K [32]. Thus, the

AN

salting-out effect by the copper sulfate is stronger than the effect by the ammonium perrhenate, since the free energies of hydration for the NH4ReO4 are: ΔG°hyd(NH4+) = −285 kJ∙mol−1 and ΔG°hyd(ReO4−) = −234

M

kJ·mol−1 at 298.15 K, although the last value is still unclear [16]. However, in the present work, from the

ED

obtained results is probable to infer that ReO4− anion has a value less negative than SO42− anion, because the

distribution increases.

PT

perrhenate anion partitioning to the top phase and with increasing salt (CuSO4) concentration the coefficient

CE

The chemical analyses show that NH4+ ion partitioning to the bottom phase, probably because the free

AC

energy of hydration is higher than the ReO4− anion.

On the other hand, the proportionality between distribution coefficient and PEG concentration can also be analyzed. Pascuali et al .[33] say that amphiphilic substances such as soap at low concentrations behave as ordinary salts, ionizing in alkali metal cation and anion of the fatty acid. But at high concentrations the anions are added together, these aggregates are called ionic micelles, in which the hydrocarbon part (lipophilic chains) of the anion is in the interior and the polar (or hydrophilic) part outwards, a micelle has an aggregation number which refers to the number of lipophilic chains. 16

ACCEPTED MANUSCRIPT Polymers are considered polyelectrolytes bearing ionizable groups, which, in polar solvents can dissociate into charged polymer chains (macroions) and small counterions [34].

Alexandridis et al. [35] studied the Poly(ethylene oxide)-poly(propylene oxide )-poly (ethylene oxide) block copolymer surfactants in aqueous solutions and in the interfaces. They found that micelles,

T

thermodynamically stable are formed with increasing copolymer concentration and/or solution temperature,

IP

as revealed by surface tension, light scattering, and dye solubilization experiments; but the aggregation

CR

number is independent of the copolymer concentration and increase with temperature.

US

Yongqiang et al. reported for the selective extraction and separation of molybdenum (VI) from aqueous media by APTS formed by Triton X-100 + (NH4)2SO4 + H2O [6], that the polymolybdate anion increases into

AN

TX-100 micellar phase, because there is a greater amount of micellar and a higher solubility with increasing

M

TX-100 concentration. Alexandridis et al. [35] and Yongqiang et al. [5] reported that the number of TX-100 molecule in each aggregate is independent of the TX-100 concentration, while the number of aggregates is

ED

proportional to the TX-100 concentration.Wu et al. [30] for the extraction and separation of butyric acid from

PT

clostridium tyrobutyricum fermentation broth in PEG/Na2SO4 aqueous two-phase system, shown that the highest yield for the separation of butyric is at 25% mass of PEG in the range 15 - 35% mass of PEG 6000. The

CE

authors concluded that the proportionality between yield production and PEG concentration is due to the

AC

formation of hydrogen bonds between hydroxyl groups of butyric acid and PEG, increasing the hydrophobicity of ATPS. The decrease of yield with increasing PEG concentration is due to saturation of PEG in the PEG-rich phase. This led to the accumulation of the excessive PEG and therefore also of butyric acid in the sodium sulfate-rich phase, resulting in a contemporary loss of yield in the PEG-rich phase. For this study and in accordance with the above considerations, with increasing PEG concentration, the amount of aggregates increases in the upper phase and due to the hydrogen bonds between hydroxyl groups of PEG

17

ACCEPTED MANUSCRIPT and the ReO4- anions, the hydrophobicity increases too, thus, the distribution coefficient for the perrhenate anion increases.

If a salting-out effect for de perrhenate anion exists, also an interaction with the PEG macromolecule should occurs, otherwise phase separation could happen only in system conformed by water and the salts, CuSO4

CR

IP

this molecule remains positively charged, so the ReO4- anion is attracted.

T

and NH4ReO4. The interaction is probably of electrostatic type, due to the protonation of the PEG molecule,

3.4 Analysis of the experimental design

US

3.4.1 Model fitting and validation

AN

The statistical significance of the determined models was evaluated by the analysis of variance (ANOVA) and least squares techniques. The ANOVA consists in determining which of the factors significantly affect

M

the response variables in study [20], using a Fisher’s statistical test (F-test). The magnitude and the

ED

significance of the effects estimates of each variable and all their possible interactions on the response variables were obtained. The effect estimate of each variable represents the improvement in the response

PT

variable, i.e. to expect as the variable setting is changed from low to high. Effects with less than 95% of

CE

significance, i.e. effects with a p-value higher than 0.05, were discarded and grouped into the error term (residual error) and for the reduced model a new analysis of variance was performed. P-value represents a

AC

decreasing index of the reliability of a result.

The regression equations were submitted to the F-test to determine the determination coefficient R2, which is defined as the ratio of the explained variation to the total variation and used to measure the degree of fitness [36] and was calculated to be 0.930 (Table 2) for distribution coefficient. This implied that 93.0% were compatible with the data predicted by the model. The R 2-value is always between 0 and 1, and a value >0.75 indicates a good response between model prediction and experimental data [17]. Although, Joglekar and May [37] suggested that R2 should be at least 0.80 for a good fitness of the model. 18

ACCEPTED MANUSCRIPT The results of the lineal model in the form of analysis of variance (ANOVA) are realized. The analysis of data was generated using Design Expert software version 7.0. Table 3 lists the statistical test results of the model. The F-values and “probability > F”-values of the two regression equations show that the model is significant. The preserved model terms were found to be significant according to the P-values; when a factor has a p-

T

value less than 0.05 influences the process in a significant way for a confidence level of 0.95. The value of the

IP

adjusted determination coefficient (adj. R2) was calculated to be 0.93 (Table 2Error! Reference source not

CR

found.) for the distribution coefficient model, which indicated a high significance of model (equation 1). All of the above considerations indicate an excellent adequacy of the lineal models to the experimental data,

US

therefore statistically valid (Figure 4). In the case of the lack of fit, the F-value is determined using the mean

AN

square of the pure error instead of the residual error. This lack-of-fit test compares the residual error, i.e. the error associated with the fitted model, to the pure error from replication. If residual error significantly

M

exceeds pure error, the model will show significant lack of fit, and another model may be more appropriate.

ED

For this work, lacks of fit results were not significant.

The results indicated that the factors of PEG 4000 concentration, CuSO 4 concentration and temperature have

PT

the greatest effect on the ReO4- anion partitioning. The PEG concentration and its interaction with system .

CE

temperature also have a significant effect on KReO4-.

AC

Table 2. Summarized data of analysis of variance (ANOVA) of the empirical model obtained for distribution coefficient according to experimental design in Table 1.

Statistical parameter Value Std. Dev.

0.31

Mean

5.99

C.V. %

5.19

19

ACCEPTED MANUSCRIPT PRESS

3.66 0.9296

Adj R-Squared

0.9191

Pred R-Squared

0.9010

Adeq Precision

23.318

IP

T

R-Squared

CR

Table 3. Summarized data of analysis of variance (ANOVA) of the empirical models obtained for

US

distribution coefficient according to experimental design in Table 1.

AN

Sum of Squares df Mean Square F Value p-value Prob > F

Model

34.4216

4

A-A-Temperature/ °C

1.8145

M

Source

89.0631

< 0.0001

1.8145

18.7796

0.0002

C-CuSO4, w/w

23.5298

1

23.5298

243.5260

< 0.0001

D-PEG 4000, w/w

8.6945

1

8.6945

89.9848

< 0.0001

0.3828

1

0.3828

3.9620

0.0568

2.6088

27

0.0966

Lack of Fit

1.2605

11

0.1146

1.3598

0.2799

Pure Error

1.3483

16

0.0843

Cor Total

37.0304

31

AC

Residual

ED

PT

CE

AD

1

8.6054

20

Significant

not significant

ACCEPTED MANUSCRIPT The statistical model in terms of coded factors:

(2) The statistical model in terms of actual factors:

(

( )

Design-Expert® Sof tware K Partition coef f icient

(3)

Predicted vs. Actual 7.90

US

Color points by v alue of K Partition coef f icient: 7.85 6.98

4.25

6.05

M

5.13

AN

Predicted

)

IP

)

)

CR

(

(

T

( )

4.25

5.15

6.05

6.95

7.85

Actual

PT

ED

4.20

CE

Figure 4. Relationship between the calculated and experimental data of coefficient distribution of ReO4- in top phase.

AC

The residues were also examined for normal distribution. Figure 5 shows the normal probability plot of residual values. It could be seen that the experimental points were reasonably aligned, suggesting normal distribution.

21

ACCEPTED MANUSCRIPT Normal Plot of Residuals

Design-Expert® Sof tware K Partition coef f icient Color points by v alue of K Partition coef f icient: 7.85

99

95 90 80 70 50 30 20 10

T

Normal % Probability

4.25

1

-0.72

0.29

1.30

CR

-1.73

IP

5

2.31

US

Internally Studentized Residuals

AN

Figure 5. Normal probability plot of residual values for the distribution coefficient of ReO 4- in top phase.

The view cube shows how three factors combine to affect the response and the optimum level of each

M

variable for maximum response, Figure 6. A high similarity between the predicted and experimental results

ED

is also observed in Figure 4, reflecting the suitability of the model to predict the desired responses.

Cube

Design-Expert® Sof tware

PT

K Partition coef f icient

CE

K Partition coef f icient X1 = A: Temperature °C X2 = C: CuSO4, w/w X3 = D: PEG 4000, w/w

7.72

7.02

C+: 0.13

6.46

6.20

C: CuSO4, w/w

AC

Actual Factor B: pH = 1.50

6.00

5.31

D+: 0.14

D: PEG 4000, w/w

C-: 0.11 D-: 0.11 4.74 4.48 A-: 35 A+: 45 A: Temperature °C

Figure 6. The “View of the cube” obtained by statistical software shows the maximum distribution coefficient (7.72) predicted by the model at w1 = 0.13, w2 = 0.137 and 35 °C. 22

ACCEPTED MANUSCRIPT Response contours are calculated by using the model equation. Contour lines as a function of PEG 4000 and CuSO4 concentration (which has greater effect) are calculated at 35°C and presented in Figure 7. The red area is a preferred experimental space with a relatively high distribution coefficient.

K Partition coefficient

Design-Expert® Software 0.13

K Partition coefficient 7.85

T

7.22

IP

4.25 0.13

CR

6.73

6.23

0.12

US

Actual Factors A: Temperature °C = 35.00 B: pH = 1.50

C: CuSO4, w/w

X1 = D: PEG 4000, w/w X2 = C: CuSO4, w/w

5.73

0.12

0.11

0.12

M

0.11

AN

5.24

0.12

0.13

0.14

ED

D: PEG 4000, w/w

Figure 7. Contour plots (2-D) showing the effects of variables, X1: PEG4000 mass fraction; X2: CuSO4 mass

PT

fraction; at 35°C and pH 2 on the response K.

CE

Thus, the anion rhenium can be separated from solutions with high SO 4- concentration, which is convenient

AC

due to the molybdenite (principal source) is usually associated with porphyry copper deposits [4].

3.4.2 Analysis of the properties of the upper and bottom phases Table 1 shows the total of experiments, from these, there are 16 cases that are unique, i.e. without repetitions. Table 4. Physicochemical properties for the bottom and upper phases at different combination of factors, temperature, T, pH, CuSO4 mass fraction,w1, PEG 4000 mass fraction, w2 and response variables, bottom phase density (ρbp), upper phase density (ρup), sound velocity of bottom phase (sbp), sound velocity of upper phase (sup), kinematic viscosity of bottom phase (νbp) and kinematic viscosity of upper phase (νup) shows these

23

ACCEPTED MANUSCRIPT cases with their response variables averaged. There are two cases in which a single parameter varies,

T

keeping the remaining variables constant. Thus, it is possible to observe the effect of parameter separately.

IP

Table 4. Physicochemical properties for the bottom and upper phases at different combination of factors,

CR

temperature, T, pH, CuSO4 mass fraction,w1, PEG 4000 mass fraction, w2 and response variables, bottom phase density (ρbp), upper phase density (ρup), sound velocity of bottom phase (sbp), sound velocity of upper

w1

w2

ρbp/g·cm-3 ρup/g·cm-3 sbp/m·s-1 sup/m·s-1 νbp/mm2·s-1 νup /mm2·s-1

45

1

0.112 0.108 1.17358

2

35

1

0.130 0.137 1.22978

1.09144

1599.72

1648.10

1.05260

11.04379

1.10389

1619.98

1676.66

1.50564

21.08642

3

35

2

0.112 0.137 1.20042

1.09023

1607.49

1667.33

1.38558

16.40415

4

35

1

5

35

2

0.112 0.108 1.18188

1.10152

1594.02

1648.27

1.30924

12.35313

0.130 0.137 1.22876

1.09807

1621.72

1678.66

1.49002

22.09172

6

35

2

7

35

1

0.112 0.108 1.18017

1.09082

1596.16

1653.32

1.29832

13.01827

0.130 0.108 1.20878

1.10280

1607.54

1664.11

1.40675

16.77075

8

45

9

45

2

0.112 0.108 1.17430

1.08939

1602.45

1655.13

1.05284

12.42254

2

0.130 0.137 1.21743

1.08703

1624.20

1670.03

1.19310

18.31415

10 11

35

1

0.112 0.137 1.20266

1.09910

1605.47

1663.43

1.36791

16.31822

45

2

0.112 0.137 1.18991

1.08483

1610.80

1661.16

1.09586

14.51172

12

45

13

35

14

45

CE

PT

ED

M

1

AN

Test T/°C pH

US

phase (sup), kinematic viscosity of bottom phase (νbp) and kinematic viscosity of upper phase (νup)

0.130 0.137 1.21930

1.09704

1622.20

1668.63

1.20610

17.36047

2

0.130 0.108 1.20599

1.09400

1609.66

1667.04

1.38515

17.47683

2

0.130 0.108 1.19752

1.08945

1613.20

1661.68

1.13823

16.06884

AC

1

15

45

1

0.130 0.108 1.19956

1.09646

1611.59

1659.81

1.18855

15.55947

16

45

1

0.112 0.137 1.19191

1.09263

1608.79

1658.25

1.10163

13.34421

The standard uncertainties u, are u(T) = 0.01 K for the density and sound velocity, u(T) = 0.05 K for the viscosity, u(ρbp) = 8·10-5 g·cm-3, u(ρup)=9·10-5 g·cm-3, u(sbp) = 0.06 m·s-1, u(sup) = 0.08 m·s-1, and u(νbp) and u(νup) = 1·10-5 mm2∙s-1.

For both phases, only density is clearly inversely proportional to pH (Figure 8). Sound velocity and kinematic viscosity do not. However, the tendency in all cases is clear.

24

ACCEPTED MANUSCRIPT

1.11 Tests 1 and 8 Tests 2 and 5

1.10

Tests 3 and10 Tests 4 and 6 Tests 7 and 13 Tests 9 and 12 Tests 11 and 16

T

1.09

Tests 14 and 15

IP

ρup/ g∙cm-3

1.10

1.08 0.7

1.2

1.7

2.2

2.7

3.2

US

pH

CR

1.09

AN

Figure 8. Effect of pH on upper phase density, ρup.

Tests 1 and 4 Tests 2 and 12

PT

1,630

ED

velocity of bottom phase (Figure 9).

M

All properties for both phases are inversely proportional to temperature, with exception of the sound

1,625

Tests 3 and 11 Tests 5 and 9

sbp/m∙s-1

AC

CE

1,620

Tests 6 and 8

1,615

Tests 7 and 15 Tests 10 and 16

1,610

Tests 13 and 14 1,605 1,600 1,595 1,590 33

38

43

48

53

T/°C

Figure 9. Effect of temperature on sound velocity of the bottom phase, sbp. 25

ACCEPTED MANUSCRIPT On the other hand, for both phases, all properties have direct proportionality with PEG and CuSO 4 concentration in both phases. Figure 10 and Figure 11 show this behavior.

1,630

Tests 1 and 15 Tests 2 and 10

1,625

T

Tests 3 and 5 1,620

IP

Tests 4 and 7

Tests 6 and 13 Tests 8 and 14

1,610

CR

sbp/m∙s-1

1,615

Tests 9 and 11

1,605

US

Tests 12 and 16

1,600

1,590 0.11

AN

1,595

0.12

0.13

0.14

M

w1

AC

CE

PT

ED

Figure 10. Effect of CuSO4 mass fraction, w1, on sound velocity of the bottom phase, sbp.

26

ACCEPTED MANUSCRIPT

23.00

Tests 1 and 16

21.00

Tests 2 and 7 Tests 3 and 6 Tests 4 and 10 Tests 5 and 13

17.00

Tests 8 and 11

T

15.00

Tests 9 and 14

IP

νup, mm2∙s-1

19.00

13.00

CR

Tests 12 and 15

11.00 9.00 0.12

0.14

US

0.1

0.16

AN

w2

M

Figure 11. Effect of PEG 4000 mass fraction, w2, on viscosity of the upper phase, νup.

Figure 11 and Figure 12 show that with increasing PEG concentration, the increase of kinematic viscosity on

ED

upper phase is more than the kinematic viscosity on bottom phase, similar behavior was found in systems

PT

PEG/Dextran [38]. In these systems, the solution viscosity of a single polymer will increase strongly with the concentration and molecular weight. Thus, the increase in viscosity provided by increasing PEG

CE

concentration is compensated by the reduction in viscosity due to the reduction in dextran concentration as

AC

more concentrated systems are considered. In the case of high molecular weight dextran/PEG systems, the lower phase viscosity is also independent of PEG molecular weight.

27

ACCEPTED MANUSCRIPT

1.60

Tests 1 and 16

1.50

Tests 2 and 7 Tests 3 and 6 Tests 4 and 10 Tests 5 and 13

1.30

Tests 8 and 11

T

1.20

Tests 9 and 14

IP

νbp/mm2∙s-1

1.40

1.10 1.00 0.90 0.12

0.14

US

0.1

CR

Tests 12 and 15

0.16

AN

w2

ED

M

Figure 12. Effect of PEG 4000 mass fraction, w2, on viscosity of the bottom phase, νbp.

The difference of properties between both phases is very important, because the phase separation or settling

PT

velocity under gravity will depend on phase-density differences or phase-viscosities. These parameters are also related with the settling area. For small density differences and high viscosity phases, long separation

CE

times are required [38]. Any transport process involving or taking place in a two phase system will depend

AC

on the viscosities of the phases.

Figure 13 and Figure 14 show a proportional behavior between density differences (between upper and bottom phases) and distribution coefficient “KReO4-”. Similar behavior for the viscosities and distribution coefficient is observed.

28

ACCEPTED MANUSCRIPT

8.00 7.50

K

7.00 6.50

T

6.00

IP

5.50

4.50

0.09

0.11

0.13

0.15

US

4.00 0.07

CR

5.00

Density differences/g·cm-3

M

AN

Figure 13. Effect of density differences of phases on distribution coefficient, KReO4-.

7.50 7.00

6.00

CE

K

PT

6.50

ED

8.00

5.50

AC

5.00 4.50 4.00 6.00

11.00

16.00

21.00

Kinematic viscosity differences/mm2·s-1 Figure 14. Effect of viscosity differences of phases on distribution coefficient, KReO4-

De Lemos et al. [29] reported that the difference in the intensive thermodynamic properties is enhanced with an increase in the TLL. According to the literature, the uneven solute partitioning will increase with 29

ACCEPTED MANUSCRIPT increasing tie-line length [39]. Therefore, the density differences are important for the partitioning. Figure 14 shows that with increasing density differences the distribution coefficient increases. With experimental data, the software design expert provided a statistical model for density differences (equation 2). The ANOVA and statistical parameters are presented in the Table 5 and Table 6, and the

)

(4)

CR

(

IP

T

following statistical model with the parameters that have more effect on density differences was obtained.

Table 5. Summarized data of analysis of variance (ANOVA) of the empirical models obtained for

AN

US

distribution coefficient according to experimental design in Table 1.

Sum of Squares

F

p-value

Square

Value

Prob > F

B-pH

3.436E-04

C-Conc SAL

4.012E-03

3

2.487E-03

434.64

< 0.0001

ED

7.461E-03

3.436E-04

60.05

< 0.0001

1

4.012E-03

701.16

< 0.0001

1

3.106E-03

542.70

< 0.0001

1.72

0.1539

CE

1

PT

Model

3.106E-03

AC

D-Conc PEG

Mean df

M

Source

Residual

1.602E-04

28

5.722E-06

Lack of Fit

9.028E-05

12

7.523E-06

Pure Error

6.995E-05

16

4.372E-06

Cor Total

7.622E-03

31

30

significant

not significant

ACCEPTED MANUSCRIPT Table 6. Summarized data of analysis of variance (ANOVA) of the empirical model obtained for density differences according to experimental design in Table 1.

Statistical parameter

0.10607

C.V. %

2.25525

PRESS

2.093E-04

CR

Mean

T

2.392E-03

IP

Std. Dev.

Value

0.9790

Adj R-Squared

AN

Pred R-Squared

US

R-Squared

0.9725

57.5244

M

Adeq Precision

0.9767

ED

4. Conclusions

This study shows the influence of system pH (Range 1–2), PEG 4000 concentration (Range 0.112-0.13), CuSO4

PT

salt concentration (Range 0.108-0.137) and temperature (35–45◦C) on the ReO4- anion distribution coefficient

CE

in the PEG/CuSO4 ATPS by using a full factorial design 2 k. The experimental design allowed the definition of appropriate model for the partitioning parameters, which in turn led to the definition of the most favorable

AC

operating conditions. The maximum distribution coefficient predicted by the model for ReO4- anion was 7.72 ± 0.31 at composition of ternary ATPS 0.137 PEG, 0.13 CuSO4 (and 0.733 H2O), all in mass fraction at 35°C, with a system pH less than 3. It is concluded that the partition coefficients of the ReO 4- anion increases with increasing PEG 4000 concentration, CuSO4 concentration and the system pH; but decreases with increasing temperature. The system pH and the temperature do not have a great effect on the partition coefficient compared with the PEG and CuSO4 concentration, especially the pH. For this reason, the pH was not considered in the model. 31

ACCEPTED MANUSCRIPT According to the results, it is concluded that the chaotropic perrhenate anion can be separated from solutions with high SO42- concentration. Thus, the extraction of rhenium from dissolved sulfate ores will increases at high sulfate concentrations.

T

ACKNOLEDGMENT

AC

CE

PT

ED

M

AN

US

CR

also to the Universidad de Antofagasta and CICITEM for the support.

IP

The authors thank CONICYT-Chile for financing this research through FONDECYT Project N 11130012, and

32

ACCEPTED MANUSCRIPT REFERENCES

[1] J.M. Casas, E. Sepúlveda, L. Bravo, L. Cifuentes, Crystallization of sodium perrhenate from NaReO4– H2O–C2H5OH solutions at 298 K, Hydrometallurgy, 113–114 (2012) 192-194. [2] COCHILCO, Mercado del Renio y su Producción en Chile, DEPP 14, 2016.

T

[3] M. Vosough, N. Shahtahmasebi, M. Behdani, Recovery Rhenium from roasted dust through super Para-

IP

magnetic Nano-particles, International Journal of Refractory Metals and Hard Materials, 60 (2016) 125-130.

CR

[4] A. Khoshnevisan, H. Yoozbashizadeh, M. Mohammadi, A. Abazarpoor, M. Maarefvand, Separation of

US

rhenium and molybdenum from molybdenite leach liquor by the solvent extraction method, Minerals & Metallurgical Processing Journal, 30 (2013) 53-58.

AN

[5] R.D. Rogers, A.H. Bond, C.B. Bauer, J. Zhang, S.T. Griffin, Metal ion separations in polyethylene glycolbased aqueous biphasic systems: correlation of partitioning behavior with available thermodynamic

M

hydration data, Journal of Chromatography B: Biomedical Sciences and Applications, 680 (1996) 221-229.

ED

[6] Z. Yongqiang, S. Tichang, H. Qingxia, G. Qing, L. Tieqiang, G. Yingchao, Y. Chunhuan, A Green Method for Extracting Molybdenum (VI) from Aqueous Solution with Aqueous Two-phase System Without any

PT

Extractant, Separation and Purification Technology, (2016).

CE

[7] P.-Å. Albertsson, Partition of cell particles and macromolecules: separation and purification of biomolecules, cell organelles, membranes, and cells in aqueous polymer two-phase systems and their use

AC

in biochemical analysis and biotechnology, Wiley New York etc.1986. [8] J. Huddleston, A. Veide, K. Köhler, J. Flanagan, S.-O. Enfors, A. Lyddiatt, The molecular basis of partitioning in aqueous two-phase systems, Trends in biotechnology, 9 (1991) 381-388. [9] J.C. Marcos, L.P. Fonseca, M.T. Ramalho, J. Cabral, Application of surface response analysis to the optimization of penicillin acylase purification in aqueous two-phase systems, Enzyme and Microbial Technology, 31 (2002) 1006-1014.

33

ACCEPTED MANUSCRIPT [10] A. Glyk, D. Solle, T. Scheper, S. Beutel, Optimization of PEG–salt aqueous two-phase systems by design of experiments, Chemometrics and Intelligent Laboratory Systems, 149 (2015) 12-21. [11] Z. Abisheva, A. Zagorodnyaya, N. Bekturganov, Review of technologies for rhenium recovery from mineral raw materials in Kazakhstan, Hydrometallurgy, 109 (2011) 1-8.

T

[12] N. Iatsenko Gerhardt, A. Palant, S. Dungan, Extraction of tungsten (VI), molybdenum (VI) and rhenium

IP

(VII) by diisododecylamine, Hydrometallurgy, 55 (2000) 1-15.

CR

[13] Y. Barrueto, K. Muñoz, K. Garnica, I. Villalobos, Y.P. Jimenez, Aqueous two-phase system of poly (ethylene glycol) 4000 and sodium molybdate at different temperatures, Fluid Phase Equilibria, 399 (2015)

US

66-73.

AN

[14] Y.P. Jimenez, H.R. Galleguillos, M. Claros, Liquid-liquid partition of perchlorate ion in the aqueous twophase system formed by NaNO 3+ Poly (ethylene glycol)+ H 2 O, Fluid Phase Equilibria, 421 (2016) 93-103.

M

[15] M. Claros, M.E. Taboada, H.R. Galleguillos, Y.P. Jimenez, Liquid–liquid equilibrium of the CuSO 4+ PEG

ED

4000+ H 2 O system at different temperatures, Fluid Phase Equilibria, 363 (2014) 199-206. [16] S.K. Spear, S.T. Griffin, J.G. Huddleston, R.D. Rogers, Radiopharmaceutical and hydrometallurgical

PT

separations of perrhenate using aqueous biphasic systems and the analogous aqueous biphasic extraction chromatographic resins, Ind Eng Chem Res, 39 (2000) 3173-3180.

CE

[17] R. Kammoun, H. Chouayekh, H. Abid, B. Naili, S. Bejar, Purification of CBS 819.72 α-amylase by aqueous

AC

two-phase systems: Modelling using Response Surface Methodology, Biochemical Engineering Journal, 46 (2009) 306-312.

[18] S. Gautam, L. Simon, Partitioning of β-glucosidase from Trichoderma reesei in poly(ethylene glycol) and potassium phosphate aqueous two-phase systems: Influence of pH and temperature, Biochemical Engineering Journal, 30 (2006) 104-108.

34

ACCEPTED MANUSCRIPT [19] K. Naganagouda, V. Mulimani, Aqueous two-phase extraction (ATPE): an attractive and economically viable technology for downstream processing of Aspergillus oryzae α-galactosidase, Process Biochemistry, 43 (2008) 1293-1299. [20] P.A.J. Rosa, A.M. Azevedo, M.R. Aires-Barros, Application of central composite design to the

T

optimisation of aqueous two-phase extraction of human antibodies, Journal of Chromatography A, 1141

IP

(2007) 50-60.

CR

[21] D.P.d. Silva, M.Z.R. Pontes, M.A.d. Souza, M. Vitolo, A. Pessoa-Junior, Infleunce of pH on the partition of glucose-6-phosphate dehydrogenase and hexokinase in aqueous two-phase system, Brazilian Journal of

US

Microbiology, 33 (2002) 196-201.

AN

[22] W.C.M. de Oliveira, G.D. Rodrigues, A.B. Mageste, L.R. de Lemos, Green selective recovery of lanthanum from Ni-MH battery leachate using aqueous two-phase systems, Chemical Engineering Journal,

M

322 (2017) 346-352.

ED

[23] G.D. Rodrigues, L.R. de Lemos, L.H.M. da Silva, M.C.H. da Silva, Application of hydrophobic extractant in aqueous two-phase systems for selective extraction of cobalt, nickel and cadmium, Journal of

PT

Chromatography A, 1279 (2013) 13-19.

[24] L.H.M. da Silva, W. Loh, Calorimetric Investigation of the Formation of Aqueous Two-Phase Systems in

CE

Ternary Mixtures of Water, Poly(ethylene oxide) and Electrolytes (Or Dextran), The Journal of Physical

AC

Chemistry B, 104 (2000) 10069-10073. [25] R.M. Banik, A. Santhiagu, B. Kanari, C. Sabarinath, S.N. Upadhyay, Technological aspects of extractive fermentation using aqueous two-phase systems, World Journal of Microbiology and Biotechnology, 19 (2003) 337-348. [26] J.A. Asenjo, Separation processes in biotechnology, CRC Press1990.

35

ACCEPTED MANUSCRIPT [27] Y. Zhang, T. Sun, Q. Hou, Q. Guo, T. Lu, Y. Guo, C. Yan, A green method for extracting molybdenum (VI) from aqueous solution with aqueous two-phase system without any extractant, Sep Purif Technol, 169 (2016) 151-157. [28] S. Raja, V.R. Murty, V. Thivaharan, V. Rajasekar, V. Ramesh, Aqueous two phase systems for the

T

recovery of biomolecules–a review, Science and Technology, 1 (2011) 7-16.

IP

[29] L.R. de Lemos, I.J.B. Santos, G.D. Rodrigues, G.M.D. Ferreira, L.H.M. da Silva, M.d.C.H. da Silva, R.M.M.

CR

de Carvalho, Phase compositions of aqueous two-phase systems formed by L35 and salts at different temperatures, Journal of Chemical & Engineering Data, 55 (2009) 1193-1199.

US

[30] W. Zhi, J. Song, F. Ouyang, J. Bi, Application of response surface methodology to the modeling of α-

AN

amylase purification by aqueous two-phase systems, Journal of biotechnology, 118 (2005) 157-165. [31] X. Wu, G. Li, H. Yang, H. Zhou, Study on extraction and separation of butyric acid from Clostridium

M

tyrobutyricum fermentation broth in PEG/Na 2 SO 4 aqueous two-phase system, Fluid Phase Equilibria, 403

ED

(2015) 36-42.

[32] Y. Marcus, Thermodynamics of solvation of ions. Part 5.—Gibbs free energy of hydration at 298.15 K,

PT

Journal of the Chemical Society, Faraday Transactions, 87 (1991) 2995-2999. [33] R.C. Pasqualí, C. Bregni, R. Serrao, Geometría de micelas y otros agregados de sustancias anfifílicas,

CE

Acta Farm Bonaerense, 24 (2005).

AC

[34] T. Radeva, Physical chemistry of polyelectrolytes, CRC Press2001. y (ethylene oxide) block

copolymer surfactants in aqueous solutions and at interfaces: thermodynamics, structure, dynamics, and modeling, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 96 (1995) 1-46. [36] G. Khayati, M. Anvari, N. Shahidi, Partitioning of β-galactosidase in aqueous two-phase systems containing polyethyleneglycol and phosphate salts, Fluid Phase Equilibria, 385 (2015) 147-152.

36

ACCEPTED MANUSCRIPT [37] A. Joglekar, A. May, Product excellence through design of experiments, Cereal foods world, 32 (1987) 857-&. [38] R. Hatti-Kaul, Aqueous two-phase systems, Aqueous Two-Phase Systems: Methods and Protocols: Methods and Protocols, (2000) 1-10.

T

[39] L.H.M. da Silva, M.C.H. da Silva, K.R. Francisco, M.V. Cardoso, L.A. Minim, J.S. Coimbra, PEO−[M (CN)

IP

5NO] x-(M= Fe, Mn, or Cr) Interaction as a Driving Force in the Partitioning of the

CR

Pentacyanonitrosylmetallate Anion in ATPS: Strong Effect of the Central Atom, The Journal of Physical

AC

CE

PT

ED

M

AN

US

Chemistry B, 112 (2008) 11669-11678.

37

ACCEPTED MANUSCRIPT Highlights The CuSO4 + PEG 4000 + H2O system was used for the partitioning of perrhenate anion.



An experimental design 2k was used for the evaluation of distribution rate.



Effect of the pH, temperature and concentration on the distribution rate was studied.



Best conditions for distribution of perrhenate anion rate were obtained.

AC

CE

PT

ED

M

AN

US

CR

IP

T



38