APPLICATION OF AN EXPERIMENTAL METHODOLOGY IN THE OPTIMIZATION OF A TUNGSTEN CONCENTRATION PROCESS BY MICROEMULSIONS

Abstract

Abstract - In this work, we applied an experimental planning methodology in order to correlate the necessary amounts with the description of the a tungsten extraction process by microemulsions. The result is a mathematical modelling carried out using the Sheffe Net method, where the mixtures concentration values are represented inside an equilateral triangle. The tungsten concentration process occurs in two stages: extraction and reextraction. The extraction stage was determined by monitoring: phase relative volume (Vr), extraction percentage (%E) and tungsten concentration in the microemulsion phase (Ct<FONT FACE="Symbol">m</font> e). The reextraction phase was determined by monitoring: reextraction percentage (%Re) and tungsten concentration in the aqueous phase (Ctaq). Finally, we obtained equations that relate the extraction / reextraction properties to the composition of specific points inside the extraction region, obeying the error limits specified for the acceptance of each parameter. The results were evaluated through the construction of isoresponse diagrams and correlation graphics between experimental values and those obtained through use of equations.

Microemulsions; experimental planning; tungsten extraction process

APPLICATION OF AN EXPERIMENTAL METHODOLOGY IN THE OPTIMIZATION OF A TUNGSTEN CONCENTRATION PROCESS BY MICROEMULSIONS

A.C.S. RAMOS, A.A. DANTAS NETO and T.N.C. DANTAS

Universidade Federal do Rio Grande do Norte - UFRN/PPGEQ - Campus Universitário - 59072.970

Natal, RN - Brazil - Phone/Fax: [phone] / E-mail: [e-mail]

(Received: September 9, 1996; Accepted: April 10, 1997)

ABSTRACT - In this work, we applied an experimental planning methodology in order to correlate the necessary amounts with the description of the a tungsten extraction process by microemulsions. The result is a mathematical modelling carried out using the Sheffe Net method, where the mixtures concentration values are represented inside an equilateral triangle. The tungsten concentration process occurs in two stages: extraction and reextraction. The extraction stage was determined by monitoring: phase relative volume (Vr), extraction percentage (%E) and tungsten concentration in the microemulsion phase (Ctm e). The reextraction phase was determined by monitoring: reextraction percentage (%Re) and tungsten concentration in the aqueous phase (Ctaq). Finally, we obtained equations that relate the extraction / reextraction properties to the composition of specific points inside the extraction region, obeying the error limits specified for the acceptance of each parameter. The results were evaluated through the construction of isoresponse diagrams and correlation graphics between experimental values and those obtained through use of equations.

KEYWORDS: Microemulsions, experimental planning, tungsten extraction process.

INTRODUCTION

A microemulsion is a liquid-liquid dispersion, formed by two liquids of different polarities (more commonly water and oil) and a mixture of surfactant agents (surfactant and/or co-surfactant). Macroscopically it exhibits a monophase liquid (homogeneous), which is translucid and isotropic appearence. Within the known proportions of its constituents this liquid is able to, equilibrate with other phases, either aqueous or organic ones. In the case of extraction by microemulsions, one can work with the equilibrium between microemulsion and the aqueous phase, which called the Winsor II system. The application of microemulsions in metal extraction is a promising and attractive process due to the acceleration of the metal adsorption kinetics, as a result of to the large increase of interfacial area in the dispersed media.

First a microemulsion diagram was defined[1]; the extraction region was delimited by a domain of interest, where the models were built for the established properties. The properties were selected in a way so as to allow a description of the entire process, by monitoring the tungsten extraction and reextraction stages.

EXPERIMENTAL METHODOLOGY

The experimental planning based on the Sheffe Net methode is applied to the mixtures [2]. In the particular case of three constituents, a series of properties varies according to the relative concentration at each point of the equilateral triangle.

Let us consider a mixture of k constituents in the proportions x1, x2, x3......xi and Y as a physical property of interest in the study. The model consists of determining a series of parameters (a1, a2, a3..), according to the polynomial degree that best matches the response (Y), until a certain specification is satisfied, which can be the difference between the calculated values and the experimental ones.

The first step consists of preparing a series of mixtures and mesuring their responses. Subsequently, we can define a region, with some points which obey a specific net, and then measure the responses at each point of the net.

We can start with a simple net, where each run consists of a relationship of the kind Y=f(xi). Ultimately, we get a system which can be solved and provides the responses in the form:

Y=å aixi + å aijxixj + å ahijxhxixj + ......... + a12...kx1x2.....xk

where: Y - response to be evaluated; ai, aij, ahij ... - parameters to be determined; xi, xj, xh...xk - composition of the specified point; k - number of constituents.

The model must be tested. The more faithfully the behavior of Y is translated according to the compositions, the better representation of the system. If this doesn’t occur, we should increase the number of points in the initial net, thus increasing the complexity of the system.

Finally, we plotted the isoresponse curves within the domain, which leads to better evaluate the property to be represented.

The mixture planning of the second type is applied to a restricted region of the equilateral triangle, which can be situated at the sides or totally inside the domain, as shown in Figure 1.

Let us consider a mixture of k constituents in real proportions x1, x2, ....xk, with the corresponding coordinates in the Scheffe Net x1’, x2’, ......xk’, represented by the triangle A’B’C’. The points situated at the vertices inside the domain of the equilateral triangle will act as pure pseudocompounds.

The relationship between the real coordinates and the ones in the model is described by the following equation:

From these coordinates, we build the experimental matrix which contains all the necessary parameters for the construction stage, tests and representation of the model.

Definition of the System

Based on previous work on tungsten extraction [3] a microemulsified system composed of by dodecilamine chloride (as surfactant), n-butanol (as co-surfactant), querosene (as oil) and sodium chloride and sodium tungstate in an aqueous solution was chosen for this study.

The process was carried out in two stages: extraction and reextraction. In the extraction stage, the aqueous solution contains sodium tungstate (0.3% close to the concentration found in the scheelita remains), and sodium chloride (0.8%, which is the necessary concentration for increasing the Winsor II region, see Figure 2, in the pseudoternary phase diagram). In the reextraction stage , we added to the system a sodium carbonate solution (8%), in order to promote the exchange between the carbonate and tungstate anions.

The process occurs in the following manner: during extraction, the tungsten in the form of the tungstate anion is adsorbed in the microemulsion phase of the Winsor II system (Figure 2) by electrostatic interaction between the polar head of the cationic surfactant and the tungstate anion, forming a complex in the microemulsion membrane. The microemulsion phase is passed to the reextraction stage, where a new solution will promote the exchange of the tungstate anions in the complex formed, allawing the recovery of the metal in a new aqueous phase at higher concentration.

Figure 1: Domain of second type, Sheffe Net.

Phase Diagram and Experimental Domain

The diagram defined for the tungsten extraction process by microemulsions is shown in Figure 3.

From the analysis of the behavior of previous parameters and from additional references that indicate that the water-rich domain is more propitious to extraction [4], we defined the triangle ACF in Figure 3 to be most suitable for tungsten extraction by microemulsions in the defined system.

Triangle ACF represents the domain (Scheffe Net) that will be explored in experimental planning.

Figure 2: Extraction and reextraction process.

Figure 3: Diagram of tungsten extraction.

RESULTS AND DISCUSSIONS

We carried out runs at specific points of the experimental region whose compositions and results (parameters to be evaluated) are described in the following matrix of experiments (Tables 1 and 2, respectively). The parameters were selected in a manner that allows a description of the entire tungsten concentration process.

Points A, B and C are employed to develop a linear model. The others are used as a test for the validity of the model. If the linear model is not acceptable, we add points D, E and F to the previous ones and develop the quadratic model. The others are used as a test for the validity of the quadratic model, and so on.

Extraction and Reextraction Models

The system was solved by using Math Cad software. The equation obtained by solving these systems was evaluated using an error analysis calculated for each parameter. The equation is acceptable for representing the parameter if it is in the range of error limits.

Extraction models

%E = 121.90x_C/T + 48.46x_a + 104.27x_o (1)

Vr = 16.50x_C/T + 1.14x_a + 8.62x_o -

- 23.57x_C/T - 17.15x_C/Tx_o - 11.47x_ax_o(2)

Ctm _e = 8072x_C/T + 7572x_a + 12560x_o -

- 29930x_C/Tx_a - 70150x_C/T -

- 41090x_ax_o + 177600x_C/Tx_ax_o(3)

Equation (1) is a linear model for the extraction percentage (%E), for a validity limit within 2%. The relative volume (Vr) is represented by a quadratic equation in (2) with a limit of + 1.8ml. The tungsten concentration in the microemulsion phase (Ctm e) is better adjusted by a reduced cubic equation (3), with a maximum error of 5% in the evaluation of this parameter.

Reextraction models

%Re = 93.89x_C/T + 102.18x_a + 94.31x_o (4)

Ct_aq = -7835.16x_C/T + 17498.12x_a - 3702.54x_o(5)

Point Mass Fractions Elasticity Module xC/T xa xo Ct(ppm) %E Vr(ml) A 0.14 0.73 0.13 2415 66 0.45 B 0.18 0.54 0.29 1185 79 1.10 C 0.21 0.34 0.45 588 89 2.68 D 0.33 0.57 0.10 1289 82 1.30 E 0.37 0.38 0.25 646 90 2.70 F 0.52 0.42 0.06 697 90 3.61 G 0.29 0.50 0.21 1220 83 1.49 H 0.30 0.56 0.14 1339 83 1.10 I 0.34 0.36 0.30 642 90 2.70 J 0.41 0.38 0.21 658 88 2.82 K 0.22 0.65 0.13 1874 77.4 0.70 L 0.20 0.43 0.37 885 87 1.60

Point	Mass Fractions			Elasticity Module
	xC/T	xa	xo	Ct(ppm) %E Vr(ml)
A	0.14	0.73	0.13	2415	66	0.45
B	0.18	0.54	0.29	1185	79	1.10
C	0.21	0.34	0.45	588	89	2.68
D	0.33	0.57	0.10	1289	82	1.30
E	0.37	0.38	0.25	646	90	2.70
F	0.52	0.42	0.06	697	90	3.61
G	0.29	0.50	0.21	1220	83	1.49
H	0.30	0.56	0.14	1339	83	1.10
I	0.34	0.36	0.30	642	90	2.70
J	0.41	0.38	0.21	658	88	2.82
K	0.22	0.65	0.13	1874	77.4	0.70
L	0.20	0.43	0.37	885	87	1.60

Table 1: of experiments for the extraction stage

Point Mass Fractions Elasticity Module xC/T xa xo CtaqII(ppm) %Re A 0.14 0.73 0.13 11195.370 100 B 0.18 0.54 0.29 6787.481 98.6 C 0.21 0.34 0.45 2637.83 96.9 D 0.33 0.57 0.10 6793.129 99.0 E 0.37 0.38 0.25 2904.010 97.1 F 0.52 0.42 0.06 3052.770 97.4 G 0.29 0.50 0.21 5546.482 98.2

Point	Mass Fractions			Elasticity Module
	xC/T	xa	xo	CtaqII(ppm) %Re
A	0.14	0.73	0.13	11195.370	100
B	0.18	0.54	0.29	6787.481	98.6
C	0.21	0.34	0.45	2637.83	96.9
D	0.33	0.57	0.10	6793.129	99.0
E	0.37	0.38	0.25	2904.010	97.1
F	0.52	0.42	0.06	3052.770	97.4
G	0.29	0.50	0.21	5546.482	98.2

Table 2: Matrix of experiments for the reextraction stage

Equations (4) and (5) show that both parameters, reextraction percentage (%Re) and tungsten concentration in the aqueous phase (Ctaq), were correlated by linear equations, with validity limits for each model at around 2% and 5%, respectively.

Correlation Curves and Isoresponse Diagrams

Figure 4 shows good correlation graphics for experimental values and those obtained by equations corresponding to each model.

Figure 4: Correlation curves for experimental and calculated data.

Figure 5 shows the behavior of studied parameters through the isoresponse diagrams obtained from the equations of the models. Analysis of these results shows that the extraction percentage becomes higher as the microemulsion phase volume increases. We can also see that there is an intense increase of the microemulsion phase volume by the addition of active matter.

Reextraction percentage and tungsten concentration in the aqueous phase increase towords the water-rich region.

%E Vr Ct_µe %Re Ct_aq Figure 5: Isoresponse Diagrams for %E, Vr , Ctm _e, %Re and Ct_aq.

CONCLUSIONS

We obtained a good correlation for the proposed models for the following parameters: extraction percentage, the relative phase volume, tungsten concentration in the microemulsion phase, reextraction percentage and tungsten concentration in the aqueous phase.

The isoresponse diagrams show that the extraction percentage becomes higher as the microemulsion phase increases, and that the reextraction percentage is more propitious in the water-rich domain. The extraction percentage is more important in the tungsten concentration process because its value depends on the phase relative volume, and this can be observed by the large range (66% to 90%) of the elasticity module. The extraction percentage is independent of volumetric factors; therefore, its response translates into a lower elasticity module.

The Sheffe Net model is a good tool for correlating parameters in extraction by microemulsions, due to the multilinear behavior that these parameters have shown within the studied domain.

At the end of the process, we noticed that the best results were obtained for a tungsten concentration up to seven times lerger than the feed concentration, where it was possible to extract up to 90% of the metal.

NOMENCLATURE

Ct Concentration of tungsten (ppm)

Ct_aq Concentration in the aqueous phase, ppm

Ct_aq II Concentration of tungsten in the aqueous phase II, ppm

Ctm _e Concentration in the microemulsion phase, ppm

E Extraction percentage, %

x_C/T Fraction Mass of co-surfactant / surfactant

x_a Fraction Mass of the aqueous phase

x_o Fraction mass of the oil phase

Vr Phase relative volume, ml

Re Reextraction percentage, %

WI, WII,

WIII and Winsor´s phases

WIV

REFERENCES

Ramos, A.C.S.; Dantas, T.N.C. and Dantas Neto, A.A., Extração de Tungstênio Utilizando Microemulsão. Anais do XVI Encontro Nacional de Tratamento de Minérios e Hidrometalurgia, Rio de Janeiro, setembro (1995).

Forestier, J.P.; Puech, E. and Tichadou, J., Application d’une Methodologie Expérimentale à l’Étude du Diagramme Ternaire d’un Gel. International J. of Cosmetic Science 7, p. 219-233 (1985).

Dantas Neto, A.A.; Dantas, T.N.C.; Duarte, M.M.L. and Avelino, S., Equilibrium Diagrams at 27 oC of the Water + Sodium Tungstate + Dodecylamine Chloride System. Journal of Chemical & Engineering Data, V38, p. 67-69 (1993).

• Ramos, A.C.S.; Dantas, T.N.C. and Dantas Neto, A.A., Extraçăo de Tungstęnio Utilizando Microemulsăo. Anais do XVI Encontro Nacional de Tratamento de Minérios e Hidrometalurgia, Rio de Janeiro, setembro (1995).
• Forestier, J.P.; Puech, E. and Tichadou, J., Application dune Methodologie Expérimentale ŕ lÉtude du Diagramme Ternaire dun Gel. International J. of Cosmetic Science 7, p. 219-233 (1985).

Publication Dates

• Publication in this collection
  09 Oct 1998
• Date of issue
  June 1997

History

• Received
  09 Sept 1996
• Accepted
  10 Apr 1997