Extraction of Bi(III) from nitrate medium by D2EHPA impregnated onto Amberlite XAD-1180

Hydrometallurgy 103 (2010) 60–67

Contents lists available at ScienceDirect

Hydrometallurgy j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / h yd r o m e t

Extraction of Bi(III) from nitrate medium by D2EHPA impregnated onto Amberlite XAD-1180 Nasr-Eddine Belkhouche ⁎, Mohamed Amine Didi Laboratory of Separation and Purification Technologies, Department of Chemistry — Faculty of Sciences, Box 119, University of Tlemcen — 13000, Algeria

a r t i c l e

i n f o

Article history: Received 5 November 2009 Received in revised form 18 February 2010 Accepted 22 February 2010 Available online 1 March 2010 Keywords: Extractant impregnated resin Bismuth (III) D2EHPA Amberlite XAD-1180 Factorial designs

a b s t r a c t Di(2-ethylhexyl)phosphoric acid (D2EHPA) was fixed on the Amberlite XAD-1180 solid support by extractant impregnated resin technique (EIR). The quantity of the adsorbed extractant in the resin was in function of the polarity of the impregnation solvent. Thus, the optimal loading of the monomeric D2EHPA was 6 g/g of dry resin using acetone as solvent. The kinetics measurements on the extraction of bismuth (III) ions from nitrate aqueous solution were conducted while varying the initial pH of aqueous solution, initial bismuth concentration, D2EHPA content in the Amberlite XAD-1180 resin, stirring speed and equilibrium time. The extraction yield of bismuth was determined at 98.5% equivalent to 490.7 mg of Bi/g of resin, so the D2EHPA–Bi complex was suggested. Also, the complete regeneration of resin was realized in two stages in the presence of concentrate hydrochloric acid. The increase of the immersion aqueous volume, slightly decreases the sorption of Bi(III). The antagonistic effect on the sorption of Bi(III) was observed by adding sodium chloride. 33 factorial designs were employed for screening the factors that would influence the overall optimization of a batch procedure of sorption. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Bismuth is used in the cosmetic products; e.g., for the preparation of creams and hair dyes, drugs manufacture; e.g., bismuth nitrate is used for treating intestinal disorders, and in other various applications (Chen et al., 2009; Shemirani et al., 2005; Thirupathi and Kim, 2009). World reserves of bismuth are usually obtained as a sub-product in lead, copper, tin and gold ores (Habashi, 2008; Jorgenson, 2003, Reyes-Aguilera et al., 2008). During the industrial metallurgical process of these ores, leaching stages with H2SO4, HCl and HNO3 are involved, and highly acidic solutions with base metals and bismuth are obtained (Donaldson and Wang, 1986; Yang et al., 2009). Bismuth is a curious metal and can be toxic in an unsuitable form (Campos et al., 2008; Habashi, 2008).The metal extraction is a major challenge in the valuable metals recovery, as well as addressing the environmental pollution problems (Campos et al., 2008; Yang et al., 2009). Conventional techniques of metal ions removal from environmental matrices include the following processes: precipitation, solvent extraction, ion exchange, adsorption, electrochemical recovery, membrane separation and other techniques that are currently the most widely used in treatment techniques (Chabani et al., 2007; Gautam and Purl, 1979; Kocaoba and Akcin, 2005). These techniques may be ineffective because they sometimes fail to meet regulation levels for technical and/or economical reasons (Belkhouche et al., 2009; Seyhan et al., 2008).

⁎ Corresponding author. Tel./fax: +213 43213198. E-mail address: [email protected] (N.-E. Belkhouche). 0304-386X/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.hydromet.2010.02.015

Solvent impregnated resin (SIR) has been postulated as an effective alternative for the separation and recovery of species from dilute solution. The use of macroporous organic polymer supports, with a high surface area and good mechanical stability, containing selective extraction reagents offers many advantages over the use of liquid– liquid extraction usually used at an industrial scale (Benamor et al., 2008; Cortina and Warshawsky, 1997). Some varieties of resin have been developed for Bi recovery, essentially for analytical purposes, more rarely for metal recovery. These resins have been essentially tested for Bi sorption at near neutral pH or in slightly acidic solutions (Campos et al., 2008). This means that the comparison of experimental data with literature data is difficult. Another resins were synthesized as supports from where, the Amberlite XAD series resins (XAD-2, XAD-16, XAD2000, etc.) have shown promise for designing chelating resins (Campos et al., 2008; Taher et al., 2004). The selection of the best extractant is highly dependent on the metal and the support used for the preparation of the extractant impregnated resin (EIR). The combination of appropriate extractant and adsorbent allows the preparation of effective and selective reactive stationary support (due to hydrophobic interaction between extractant and the polymer network), since the stability of EIR caused the leakage of the extractant from the polymeric structure (Navarro et al., 2008). The extractants most frequently (Draa et al., 2004; Reyes-Aguilera et al., 2008) are: neutral or acidic oraganophosphorus compounds such as: di(2-ethylhexyl)phosphoric acid (D2EHPA). This last is very stable at temperatures lower or equal to 60 °C. It can be used in several cycles of the extraction processes during several months without risk of decomposition. In solvent extraction processes, D2EHPA interact with metals ions by cations exchange to form the metal complex

N.-E. Belkhouche, M.A. Didi / Hydrometallurgy 103 (2010) 60–67

(Belkhouche et al., 2005). Many studies dealing with sorption and separation of metal with SIR have been carried out. These researches were mainly devoted to explain the impregnation processes, to study the physical structure of SIR beads, to elucidate the equilibrium reactions involved as well as kinetics (Navarro et al., 2008; Seyhan et al., 2008). The approaches used in studying chemical equilibrium are based on those developed for liquid–liquid extraction. Through the literature some papers were concerned with the bismuth (III) extraction from nitrate solutions by SIR using the Amberlite XAD-1180 resins as porous support with high specific area. The aim of this present work is to study the influence of experimental parameters on the kinetics of bismuth (III) extraction from nitrate solutions by SIR using the Amberlite XAD-1180 resin impregnated with D2EHPA as organophosphorus extractant, in order to know the best operating conditions of later selective extraction from other metals such as lead, copper and tin. The experimental values were used to determine the polynomial model constants which are adjusted to the studied parameter variations of bismuth (III) extraction. 2. Experimental 2.1. Reagents The Amberlite XAD-1180 resin is a polystyrene divinylbenzene copolymer support (the size of resin particles was 20 to 50 mesh; i.e., 250–710 μm) with specific area higher than 500 m2/g, it was supplied by Fluka. The porosity resin is equal to 0.6 and the porous volume was 1.1 cm3/g. The maximum temperature of work was equal to 110 °C with a loss drying of 65%. D2EHPA (Fluka) containing 40% of M2EHPA; mono(2-ethylhexyl)phosphoric acid, of analytical grade was used without further purification. The stock of pure nitrate of bismuth (Bi (NO3)3) was supplied by Reachim and potassium iodide was from Gerhard Buchmann. Heptane and sulphuric acid were supplied by Fluka where potassium hydroxide was from Merck. Acetone was provided by Rectapur™. Others reagents were analytical grade and were used without purification. 2.2. Apparatus Haier mechanical agitator with standard platform was used in extraction experiments. pH measurements were taken with a Consort C831 pH-meter using a combined electrode. The weighing was made with an electronic analytical balance; type Kern ABS. Kikawerke TC-2 hotplate agitator equipped with a standard thermocouple, was used for the realization of the temperature experiments. UV–visible absorption spectrophotometer; type Lambda 800 Perkin Elmer was used for the bismuth (III) analysis.

61

phosphoric acid (D2EHPA) impregnated in XAD-1180 resin were carried out in batch system at 22 °C. Five milliliters of the bismuth aqueous solution is preliminary mixed with 0.1 g of XAD-1180 resin impregnated with the D2EHPA, and the whole is carried under mechanical stirring. After the end of the extraction reaction, the bismuth solution was separated using filter paper. Each extraction experiment was repeated three times and then the value average is taken in calculation. The aliquots of 10 µL of bismuth aqueous solution, were taken for analysis and the bismuth concentration was determined Visible spectrophotometrically by measuring their maximum absorbance at 460 nm by iodide method (Charlot, 1978; Charlot, 1983; Jeffery et al., 1989). 3. Results and discussion 3.1. Amberlite XAD-1180 impregnation Fig. 1 shows the influence of the nature of the impregnation solvent on the quantity of adsorbed D2EHPA into Amberlite XAD-1180. At the extractant concentration lower than 0.4 mol L− 1, no notable sensitivity was observed of the impregnation toward the polarity of the impregnated solvent. Then, the extractant content varies linearly with the extractant concentration in the solution, in the case of both solvents (acetone and heptane). Indeed, when the concentration of D2EHPA is higher than 0.4 mol L− 1, the extractant retention on the XAD-1180 resin is more quantitative in the impregnation experiments using D2EHPA dissolved in acetone (hydrophilic solvent). In these conditions, a plateau value of extractant loading of 18.60 mol/kg of dry resin is reached for D2EHPA concentration equal to 0.62 mol/L in acetone as solvent (see Fig. 1). We noted that, the SIR surface form did not change after this maximal of loading. This result is allotted to the D2EHPA solubility which is generally governed by the nature of the solvent in which it dissolved (Belkhouche et al., 2005). This behavior influences the evolution of the impregnation of XAD-1180 resin by the D2EHPA. The latter is present in monomeric form in the polar solvent such as acetone (Juang and Su, 1992; Sainz-Diaz et al., 1996), which facilitates the insertion in Amberlite XAD-1180 pores. It is also noticed that, when the D2EHPA concentration in acetone is higher than 1.1 mol L− 1, the impregnation rate of the resin increases again. This is due to the aggregate formation of D2EHPA molecules on resin surface (Navarro et al., 2008). The above-saturated concentration corresponds to 6 g of D2EHPA adsorbed per gram of dry XAD-1180 resin. From the measured density of pure D2EHPA at 298 K (0.974 g/cm3) and the pore volume of this support is 1.1 cm3/g resin, we observed that this likely means that the

2.3. Preparation of impregnated resin Before impregnation, the XAD-1180 resin was washed with distilled water in order to remove inorganic impurities and monomeric material, then the resin was dried in the drying oven, at 40 °C. For the impregnation method, we have followed the dry method (Benamor et al., 2008). A quantity of 0.1 g dry XAD-1180 resin was placed in 5 mL of heptane or acetone solvent containing D2EHPA as extractant at different concentrations for 12 h, under a stirring speed of 230 rpm. The latter was selected in this study. The polymeric beads were separated through a filter paper. The extractant remaining in the solvent was determined by potentiometric titration with KOH and the D2EHPA content of the impregnated resin was calculated from mass balance. 2.4. Extraction procedure The extraction experiments of bismuth (III) starting from aqueous solution of bismuth (III) nitrate, by the impure di(2-ethylhexyl)

Fig. 1. Evolution of the Amberlite XAD-1180 physical impregnation in function of D2EHPAr concentration in resin.

62

N.-E. Belkhouche, M.A. Didi / Hydrometallurgy 103 (2010) 60–67

impregnation of impure D2EHPA on XAD-1180 is not mainly due only to pore filling but also to structural surface. This result is more advanced than the previous work devoted on impregnation of Amberlite XAD-7 by pure D2EHPA in ethanol where the loading saturation of resin is only due to pore filling (Benamor et al., 2008). Under our extraction conditions, the stability of EIR technique is kept by the passage of M2EHPA in water due to its strong solubility in the latter. This is confirmed thereafter, in the study of evaluation of D2EHPA/XAD-1180 ratio on Bi(III) sorption where the quantity of D2EHPA that remained in the resin pores gives the maximum sorption. We noted that, the extraction yield of bismuth ions by D2EHPA molecules remained practically constant for more than 6 h of extraction time, under an agitation speed of 230 rpm (Fig. 4). 3.2. Evaluation of D2EHPA/XAD-1180 ratio on Bi(III) sorption The previous results of the impregnation process study of Amberlite XAD-1180 resin showed that the quantity of the D2EHPA in the resin plays an important role in the bismuth (III) sorption. In order to investigate the effect of extractant concentration in the resin, this parameter was varied from 15.0 to 46.5 mmol/g SIR. Fig. 2 shows the variation of the extracted bismuth rate with time at different concentrations of DEHPA in the XAD-1180 resin phase. This series of experiments was carried out with initial bismuth (III) concentration equal to 250 ppm and at initial pH of 3.6. From the results, we conclude that the equilibrium sorption of bismuth (III) is reached at 30 min of agitation, whereas sorption kinetic is faster at low extractant content of the resin (Fig. 2). This puts the assumption that, only an adequate quantity which respects the pore volume is sufficient to have an optimal sorption. These results are in agreement with those carried out on XAD-2, XAD-4, XAD-16 (Matsunaga et al., 2001) or XAD-7 resins (Benamor et al., 2008). Therefore, under our operating conditions the pore volume of XAD-1180 resin corresponds to 15.0 mmol/g SIR and above this value the sorption rate is not hardly influenced by the extractant loading of the resin. 3.3. Effect of the pH on Bi(III) sorption by SIR The pH is an important parameter in the study of the sorption kinetic and equilibrium. Its importance is discussed in terms of the metal form in the aqueous solution, as well as the functional group on the extractant. The studied pH range was taken below the precipitation of metals (pH b 5). The results given by Fig. 3 show that, the extraction rate of Bi3+ is very sensitive to the change of the initial pH of the metal aqueous phase. The best extraction yield (98.5%) was obtained at pH equal to 3.6. Beyond this pH the extraction yield falls

Fig. 2. Variation of the extracted bismuth rate with time at different concentrations of D2EHPAr in the XAD-1180 resin phase.

Fig. 3. pH effect on the adsorption of Bi3+ onto SIR. Initial bismuth (III) concentration 250 ppm, time 75 min, extractant content [HR]r 15.0 mmol/g SIR.

freely. This is allotted to the increase in the acidity of the medium because the advance of the extraction reaction makes the pH decrease in the aqueous phase, and to the decrease of the concentration of free bismuth (III). From there, we will assist to a back-extraction of Bi+ 3 by the D2EHPA. This mode of extraction is due to the chemistry of D2EHPA which acts with a cation exchange mechanism (Belkhouche et al., 2005; Belkhouche et al., 2006). 3.4. Contact time and stirring speed From the results shown in Fig. 4, we note that for a given stirring speed, the extraction yield of bismuth (III) by D2EHPA/XAD-1180 resin increases according to the time. In general, it becomes constant for the times higher than 75 min. When the agitation speed reached a value of 230 rpm the concentration profile did not change any more. This was observed in the case of 275 rpm of stirring. This phenomenon indicates that the effect of the fluid to bismuth mass transfer resistance became negligible. Thus, we obtained a maximum value of 98.5% of extraction yield at 230 rpm after 30 min of stirring. Further experiments were conducted with a stirring speed of 230 rpm in order to save energy and a contact time of 75 min to avoid the concentration effects. 3.5. Study of Bi(III) sorption on SIR The study of the equilibrium for bismuth (III) sorption from aqueous solutions by solvent impregnated XAD-1180 resin was carried out by

Fig. 4. Influence of the agitation speed on the sorption kinetics. Initial bismuth (III) concentration 250 ppm, initial pH 3.6, D2EHPA content [HR]r 31.0 mmol/g SIR.

N.-E. Belkhouche, M.A. Didi / Hydrometallurgy 103 (2010) 60–67

63

Fig. 5. Extension of Bi(III) concentration on SIR sorption at different extractant contents. Time 75 min.

Fig. 6. Influence of initial concentration of Bi(III) on sorption kinetic. Extractant content [HR]r 15.0 mmol/g SIR and initial pH 3.6.

taking account the results given in process evaluation of Amberlite XAD1180 impregnation. In this context, three different SIR containing 15.0, 31.0 and 46.5 mmol DEHPA/g of XAD-1180 resin at an initial pH of 3.6 were tested at increasing initial bismuth (III) concentration until saturation of the resin. The results are presented in Fig. 5. Each of these curves show a strong increase in the sorption capacity at low residual metal concentration (lower than 5000 ppm) in particular when the D2EHPA content in XAD-1180 resin was equal to 15.0 mmol/g of resin. The maximum sorption capacity decreases with increasing extractant concentration in XAD-1180 resin, due to the destabilization of EIR during the extraction process. It varies from 490.7 to 466.4 mg of Bi/g SIR when the extractant loading varies from 15.0 to 46.5 mmol/g SIR. These results are in good agreement with those shown in the impregnation process study of XAD-1180 resin where the equilibrium of saturation of resin is reached at 0.3 mol L− 1 of D2EHPA (for 0.1 g of resin). So, the maximum SIR capacity determined is 490.7 mg of Bi/g SIR obtained for a SIR containing 15.0 mmol DEHPA/g XAD-1180 resin. At saturation the molar ratio D2EHPA/Bi is close to 6. This, suggests that 6 mol of the monomer of D2EHPA is necessary for complexes with 1 mol of bismuth from where the suggested equilibrium of the extraction reaction is given by Eq. (1) (McKevitt and Dreisinger, 2009).

EIR recycling. From the investigations made in the study of Bi(III) sorption on SIR, we note that the bismuth loaded in the XAD-1180 resin in the form of D2HPA–Bi(III) complexes can be back-extracted in a concentrate acid medium. This is kinetically realized while moving the extraction reaction (Eq. (1)) towards the left i.e., for the bismuth (III) solution concentrate. The method of back-extraction of Bi(III) from loaded XAD-1180 resin consist in destabilizing of D2HPA–Bi(III) complexes by the addition of a concentrate stripping solution of hydrochloric acid. It's possible because of the extraction mode with organophosphorus acid (Belkhouche et al., 2005). Several experiments were carried out on a XAD-1180 resin charged to the maximum with bismuth (98.5%) and containing 15.0 mmol of D2EHPA/g SIR. The results show (Fig. 7) that the bismuth is eluted from impregnated XAD-1180 resin by D2EHPA to 51%, in only one stage at a very lower pH. Therefore, two stages will be sufficient for its total desorption.

3þ

Bi

þ 6HR⇔BiR3 ·3HR þ 3H

þ

ð1Þ

3.6. Effect of concentration of Bi(III) on sorption

3.8. Effect of aqueous phase volume The study of the influence of the aqueous phase volume on sorption was studied by immersing the resin by various volumes of the bismuth (III) aqueous solution. The concentration was maintained constant and equal to 250 ppm at the optimal pH of 3.6. The results show (Fig. 8) that the best extraction yield of bismuth (III); 98.5% was obtained by immersing XAD-1180 resin with a volume of 5 mL. By

The study of the influence of the initial concentration of bismuth (III) on the sorption kinetic shows that by increasing the initial concentration of bismuth (III) in aqueous solution from 100 to 250 ppm, the uptake of Bi(III) becomes higher (Fig. 6). Thus, the concentration gradient occurs at the solution–sorbent interface. So, the driving force, which is the difference between the bulk phase concentration and the resin phase concentration, controlled the diffusion transport. When increasing the initial concentration to 500 ppm, no substantial difference was observed on Bi(III) sorption. This is an indication of internal diffusion processes of metal ions in the pores of resin. Also, when the initial concentration of Bi(III) varies from 100 to 500 ppm, the time; t50 for 50% attainment of equilibrium remains nearly independent of the metal initial concentration. 3.7. Elution of Bi(III) from loaded EIR The possibility of desorbing the metal from the loaded EIR is also an important criterion for the processes of a sorption system. The experiments of the adsorbed bismuth elution allow the later study of

Fig. 7. Evolution of back-extraction of Bi(III) with pH of stripping phase. Time 30 min.

64

N.-E. Belkhouche, M.A. Didi / Hydrometallurgy 103 (2010) 60–67 Table 1 Factor levels used in the 33 factorial experiment designs.

Fig. 8. Effect of aqueous phase volume on Bi(III) sorption. Time 30 min, extractant content [HR]r 15.0 mmol/g SIR.

increasing the volume of bismuth (III) 24 times, we lose 16% of extracted bismuth (III). 3.9. Effect of electrolyte The influence of the ionic strength of the aqueous phase on the bismuth (III) sorption, by the D2EHPA impregnated onto Amberlite XAD-1180, was controlled by the addition of various quantities of the sodium salt to the metal aqueous phase. From the results given in Fig. 9, we notice that the quantity of extracted bismuth (III) decreases linearly with the increase of the added NaCl quantity to the metal aqueous phase. The electrolyte NaCl exerts an antagonistic effect on sorption, an extraction yield of bismuth (III) of 98.5% (in absence of salt) to 81% in the presence of 0.06 g/L of salt (maximum of solubility). This is explained by the fact that the Na+ ion enters in competition with Bi3+ during the extraction with the D2EHPA. This is due to the physicochemical character of D2EHPA in the salted mediums (Sato and Nakamura, 1972). This result is similar to that found by other authors (Belkhouche et al., 2005) carried on the influence of the NaCl electrolyte on the liquid–liquid extraction of copper (II) by the D2EHPA. 4. Factorial design study The experiments showed that, as expected, numerous factors can influence the bismuth (III) extraction but only some of them, namely the

Fig. 9. Effect of NaCl electrolyte on Bi(III) sorption. Initial bismuth (III) concentration 250 ppm, extractant content [HR]r 15.0 mmol/g SIR, time 30 min.

Parameter level

Reduced value

X1 (g /g of resin)

X2 (ppm)

X3

Minimal Medium Maximal

−1 0 +1

2.5 5 10

90 250 700

3 3.6 4

extractant in XAD-1180 resin, initial concentration of bismuth (III) and the pH value can be regarded as being the key-parameters that govern the process efficiency. An adequate selection of these parameters is an essential requirement for establishing an accurate polynomial model (Eq. (2)). The knowledge of the variation field of each factor (minimal, medium and maximal) makes it possible to determine the exactitude limits of the experimental model in question (Sekkal et al., 2009). Thus, describes the investigated process (Assaad et al., 2007; Ferreira et al., 2006; Star et al., 2002). The limits of the variable ranges must take into account the results of the preliminary tests. In our investigations, a series of 27 attempts was made according to a 33 experiment factorial design by varying the D2EHPA/XAD resin ratio (X1), the initial concentration of bismuth (X2) and the pH value (X3) in suitable parameter ranges. Three variation levels for each parameter were considered as summarized in Table 1. Y = b0 + b1 X1 + b2 X2 + b3 X3 + b4 X1 X2 + b5 X1 X3 + b6 X2 X3 +

2 b7 X1

+

2 b8 X2

+

2 b9 X3

ð2Þ

+ b10 X1 X2 X3 :

4.1. Experiment design for bismuth ion extraction The results of the bismuth extraction process were expressed in terms of the extraction yield, regarding as being the response function in the investigated process. These results are summarized in Table 2. Preliminary observations show that the extraction yield of Bi(III) significantly Table 2 Experimental design and extraction capacity (%) of D2EHPA. Experiment no.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28, 29, 30a

Factor levels

Response function

X1

X2

X3

Extraction yield (Y %)

−1

−1

−1 0 1 −1 0 1 −1 0 1 −1 0 1 −1 0 1 −1 0 1 −1 0 1 −1 0 1 −1 0 1

97.86 96.26 95.61 95.45 96.10 96.79 96.37 95.38 97.19 94.66 92.53 92.08 88.44 90.91 91.83 93.62 85.33 89.02 90.77 92.07 95.69 91.23 94.76 96.18 90.99 93.58 92.40 95.97, 95.54, 95.12

0

+1

0

−1

0

+1

+1

−1

0

+1

a Three additional tests at the central point (0, 0, 0) for the calculation of the Student and Fisher's tests, using the normal rule of variance.

N.-E. Belkhouche, M.A. Didi / Hydrometallurgy 103 (2010) 60–67

65

Table 3 Model coefficients and their corresponding effects upon yield extraction of Bi(III). Variable Model

Expected effect on the yield extraction

Coefficient Value X0 = 1 X1

b0 b1

X2 X3 X1 X2

b2 b3 b4

X1 X3

b5

X2 X3 X21 X22 X32 X1X2X3

b6 b7 b8 b9 b10

90.45 High average extracting capacity of the D2EHPA 3.539 (+++) Very strong advantageous individual effect of X1 1.905 (++) Strong advantageous individual effect of X2 0.617 (+) Favourable individual effect of X3 −0.158 (−) Weak detrimental binary interaction of X1 and X2 0.075 (+++) Very favourable binary interaction X1 and X3 0.025 (+) Favourable binary interaction of X2 and X3 −7.672 Pronounced maximum with respect to X1 −3.005 Pronounced maximum with respect to X2 −0.239 Slight and flat maximum with respect to X3 −0.637 (−) Weak ternary detrimental interaction of X1, X2 and X3

(+) Favourable or positive effect; (−) detrimental or negative effect.

according to the experiment parameters, reaching values of 85.33–97.86% under certain operating conditions. From Table 2, it already appears that the highest yield extraction value (97.86%) was obtained for minimal D2EHPA/XAD resin ratio, minimal initial concentration of bismuth (III) and minimal pH value. 4.2. Model calculations and refinement The bismuth extraction modelling was achieved on the basis of the 27 measured values, using a Taylor's second-order polynomial (Assaad et al., 2007). The model calculations were achieved using non dimensional values of these variables. Table 3 summarizes the coefficient values of the model, supposed to describe the individual effects of parameters, along with their possible interactions. The individual effects and interactions of the parameters were discussed on the basis of the sign and the absolute value of each coefficient. These coefficient features will define the strength of the corresponding effect involved and the way it acts upon extraction yield (favourable or detrimental), respectively. The first observations from Table 3 already allow making the following statements:

Fig. 10. 3-D representations of the extraction yield of bismuth (III) at fixed: (a) D2EHPA/XAD resin = 5, (b) [Bi(III)]i = 250 ppm, (c) pH = 3.6.

i. high extracting capacity of the D2EHPA ought to be obtained within the fixed parameter ranges, thereby justifying the suitable choice of the limits; ii. the effect of the D2EHPA/XAD resin ratio is three times stronger than pH, while the initial concentration of bismuth plays an important role within the investigated ranges; iii. interaction between D2EHPA/XAD resin and the pH of bismuth aqueous solution which strongly governed the extraction; iv. except between D2EHPA/XAD resin ratio and pH, bismuth concentration and pH, all interactions are unfavourable; v. pronounced maximum with respect to D2EHPA/XAD resin and to a lesser extent to the pH of solution, will characterize the response surface, giving rise to precise optimal values of these parameters; vi. no synergy must be involved simultaneously between the three parameters.

mining which coefficients could be neglected, through Student and Fisher tests (Assaad et al., 2007; Bergouzini and Duby, 1995). The model adequacy strongly depends on the accuracy of the experiment. In the current experiment, the main errors arise from volume and weight measurements. For this purpose, three additional attempts at the central point (0, 0, 0) are required for estimating the average error in the value of each coefficient, on the basis of the random variance. The calculations made are summarized in Table 5. Thus, with a 95% of confidence (i.e., α = 0.05), and for a 2 variance (i.e., for three attempts at central point), the confidence range for all the coefficients estimated using 27 runs (N = 27), will be Δbi = ±0.351 at 95% of confidence. From the Student's test, it results that |bi| b |Δbi| for b4, b5, b6 and b9. Consequently these coefficients must be removed from

These predictions are in agreement with the shape of the response surface (Fig. 10a, b and c), plotted three times by fixing successively the three parameters at the central values, according to the equations in Table 4. The vicinity around these central values is supposed to include the optimum. For the sake of the model reproducibility, we must check whether this model accurately describes the process investigated by deter-

Table 4 Specific regression functions with one fixed variable. Fixed coded variable

Polynomial model

X1 = 0 X2 = 0 X3 = 0

Y = 90.45 + 1.905X2 + 0.617X3 + 0.025X2X3 − 3.005X22 − 0.239X23 Y = 90.45 + 3.539X1 + 0.617X3 + 0.075X1X3 − 7.672X21 − 0.239X23 Y = 90.45 + 3.539X1 + 1.905X2− 0.158X1X2 − 7.672X21 − 3.005X22

66

N.-E. Belkhouche, M.A. Didi / Hydrometallurgy 103 (2010) 60–67

Acknowledgements

Table 5 Model adequacy tests and variance analysis. Feature

Symbol

Average yield at (0,0,0) point Random variance Square root of variance Risk factor (chosen arbitrary) Student's test factor Average error on the coefficient value Number of remaining coefficients Model response at (0,0,0) Discrepancy on average yield Error on average yield discrepancy Average yield for the 27 attempts Residual variance Degrees of freedom Residual degrees of freedom Fisher's test

y0 S2 S α tv Δbi R b0(y000) d Δd Ym S2r v1 v2 F

a b c

Value 95.54 0.180 0.425 0.05 (95%) 4.23 0.351a 7b 90.45 5.086 1.112 93.48 13.29 3 6 73.61c

Student Law with 2 degrees of freedom at a 95% confidence (t2,0.975). After removing the less significant coefficients. See Fisher–Snedecor tables.

the mathematical model because they do not display significant effect upon the response function, being shaded by their average error. Consequently, the final form of the polynomial model that describes the bismuth (III) extraction by D2EHPA in XAD-1180 resin is given by Eq. (3): 2

2

Y = 90:45 + 3:539X1 + 1:905X2 + 0:617X3 −7:672X1 −3:005X2 −0:637X1 X2 X3 :

ð3Þ This model is supposed to accurately fit to the extraction process of bismuth (III) investigated herein. Thus in the vicinity of the expected optimal parameter values, it appears that only D2EHPA/XAD resin ratio and interactions between D2EHPA/XAD resin ratio and pH of aqueous solution are important effects on the Bi(III) extraction. In addition, adequacy tests were applied to check whether the model calculated is valid within the parameter ranges investigated. For this, while comparing the observed value of Fisher's factor (73.61) with the critical one, given by the tables of Fisher's and that equal to 6.60, it's noticed that the Fobserved ≫ Fcritical, indicating that the model is linear and can be applied within the whole range investigated. 5. Conclusion The investigations carried out on the sorption of Bi(III) from nitrate medium by impure D2EHPA, impregnated onto the Amberlite XAD-1180 showed that an optimal plateau value of the D2EHPA monomeric extractant loading is 6 g/g of dry resin using acetone as solvent. The maximum of sorbed bismuth is 98.5% obtained from initial concentration of Bi(III) of 250 ppm present in 5 mL of aqueous solution at initial pH equal to 3.6 under the operating conditions such as: stirring speed of 230 rpm, equilibrium time of 30 min and 15.0 mmol/g SIR. We noted that, the concentrate solution of Bi(III) obtained for sorption saturation is 5000 ppm. From the maximum SIR capacity (490.7 mg of Bi/g SIR), we suggest that 6mol of the monomer of D2EHPA is necessary for complexes with 1mol of bismuth. The elution of the adsorbed bismuth in XAD-1180 resin impregnated by D2EHPA is 51% in only one stage at lower pH. The results showed that the presence of the NaCl electrolyte in aqueous solution of bismuth exerts an antagonistic effect on the sorption. In order to achieve the best conditions for Bi(III) extraction from nitrate aqueous solution by D2EHPA impregnated in XAD-1180 resin, a full 33 factorial designs were employed for screening the factors that would influence the overall optimization of a batch procedure of sorption. This optimization showed that only D2EHPA/XAD resin ratio and interaction between D2EHPA/XAD resin ratio and pH of aqueous solution play important effects on the Bi(III) extraction.

The authors thank the University of Tlemcen and ANDRS-Algeria for the financial support. References Assaad, E., Azzouz, A., Nistor, D., Ursu, A.V., et al., 2007. Metal removal through synergic coagulation–flocculation using an optimized chitosan–monmorillonite system. Appl. Clay Sci. 37, 258–274. Belkhouche, N., Didi, M.A., Villemin, D., 2005. Separation of nickel and copper by solvent extraction using di-2-ethylhexylphosphoric acid-based synergistic mixture. Solvent Extr. Ion Exch. 23 (5), 677–693. Belkhouche, N., Didi, M.A., Romero, R., Jonsson, J.Åke., Villemin, D., 2006. Study of new organophosphorus derivates carriers on the selective recovery of M (II) and M (III) metals using supported liquid membrane extraction. J. Membr. Sci. 284, 398–405. Belkhouche, N., Didi, M.A., Taha, S., Farès, N.B., 2009. Zinc rejection from leachate solutions of industrial solid waste— effects of pressure and concentration on nanofiltration membrane performance. Desalination 239, 58–65. Benamor, M., Bouariche, Z., Belaid, T., Draa, M.T., 2008. Kinetic studies on cadmium ions by Amberlite XAD7 impregnated resins containing di(2-ethylhexyl) phosphoric acid as extractant. Sep. Purif. Technol. 59, 74–84. Bergouzini, J.C., Duby, C., 1995. Analyse et planification des expériences, Les dispositifs en blocs. Masson, Paris, France. Campos, K., Domingo, R., Vincent, T., Ruiz, M., Sastre, A.M., Guibal, E., 2008. Bismuth recovery from acidic solutions using Cyphos IL-101 immobilized in a composite biopolymer matrix. Water Res. 42, 4019–4031. Chabani, M., Amrane, A., Bensmaili, A., 2007. Kinetics of nitrates adsorption on Amberlite IRA-400 resin. Desalination 206, 560–567. Charlot, G., 1978. Dosages aborptiométriques des élements minéraux, p. 182. Charlot, G., 1983. Les réactions chimiques en solution aqueuse et caractérisation des ions, p. 257. Chen, H.Z., Kao, M.C., Young, S.L., Yu, C.C., Lin, C.H., Lee, C.M., Ou, C.R., 2009. Effects of annealing atmosphere on microstructure and ferroelectric properties of praseodymium-doped Bi4Ti3O12 thin films prepared by sol–gel method. Thin Solid Films 517 (17), 4818–4821. Cortina, J.L., Warshawsky, A., 1997. Developments in solid–liquid extraction by solvent impregnated resins. In: Marinsky, J.A., Marcus, Y. (Eds.), Ion Exchange and Solvent Extraction, vol. 13. Marcel Dekker, New York, p. 195. Donaldson, E.M., Wang, M., 1986. Determination of silver, antimony, bismuth, copper, cadmium and indium in ores, concentrates and related materials by atomicabsorption spectrophotometry after methyl isobutyl ketone extraction as iodides. Talanta 33 (3), 233–242. Draa, M.T., Belaid, T., Benamor, M., 2004. Extraction of Pb(II) by XAD7 impregnated resins with organophosphorus extractants (DEHPA, IONQUEST 801, CYANEX 272). Sep. Purif. Technol. 40, 77–86. Ferreira, H.S., Bezerra, M.A., Ferreira, S.L.C., 2006. A pre-concentration procedure using cloud point extraction for the determination of uranium in natural water. Microchim. Acta 154, 163–167. Gautam, M., Purl, B.K., 1979. Spectrophotometric determination of bismuth, cobalt, and nickel after extraction of their xanthates with molten naphthalene. Mikrochim. Acta I 515–523. Habashi, F., 2008. Arsenic, antimony, and bismuth production. Encyclopedia of Materials: Science and Technology, pp. 332–336. Jeffery, G.H., Bassett, J., Mendham, J., Denny, R.C., 1989. Textbook Quantitative Chemical Analysis, Fifth Edition. VOGEL'S, p. 684. Jorgenson, J.D., 2003. U.S Geological Survey. Mineral Commodity Summaries, p. 37. Juang, R.S., Su, J.Y., 1992. Sorption of copper and zinc from aqueous sulfate solutions with bis(2-ethylhexyl)phosphoric acid-impregnated macroporous resin. Ind. Eng. Chem. Res. 31 (12), 2774–2779. Kocaoba, S., Akcin, G., 2005. Removal of chromium (III) and cadmium (II) from aqueous solutions. Desalination 180, 151–156. Matsunaga, H., Ismail, A., Wakui, Y., Yokoyama, T., 2001. Extraction of rare earth elements with 2-ethylhexyl hydrogen 2-ethylhexyl phosphonate impregnated resins having different morphology and reagent content. React. Funct. Polym. 49 (3), 189–195. McKevitt, B., Dreisinger, D., 2009. A comparison of various ion exchange resins for the removal of ferric ions from copper electrowinning electrolyte solutions Part II: electrolytes containing antimony and bismuth. Hydrometallurgy 98 (1–2), 122–127. Navarro, R., Saucedo, I., Núñez, A., Ávila, M., Guibal, E., 2008. Cadmium extraction from hydrochloric acid solutions using Amberlite XAD-7 impregnated with Cyanex 921 (tri-octyl phosphine oxide). React. Funct. Polym. 68 (2), 557–571. Reyes-Aguilera, J.A., Gonzalez, M.P., Navarro, R., Saucedo, T.I., Avila-Rodriguez, M., 2008. Supported liquid membranes (SLM) for recovery of bismuth from aqueous solutions. J. Membr. Sci. 310, 13–19. Sainz-Diaz, C.I., Klocker, H., Marr, R., Bart, H.J., 1996. New approach in the modelling of the extraction equilibrium of zinc with bis-(2-ethylhexyl) phosphoric acid. Hydrometallurgy 42 (1), 1–11. Sato, T., Nakamura, T., 1972. The complexes formed in the divalent transition metal– sulphuric acid-di-(2-ethylhexyl)-phosphoric acid extraction systems—cobalt(II), nickel(II) and copper(II) complexes. J. Inorg. Nucl. Chem. 34, 3721–3730. Sekkal, A.R., Didi, M.A., Belkhouche, N., Canselier, J.P., 2009. Removal of chromium (III) by two-aqueous phases extraction. J. Hazard. Mater. 167 (1–3), 896–903. Seyhan, S., Merdivan, M., Demirel, N., 2008. Use of o-phenylene dioxydiacetic acid impregnated in Amberlite XAD resin for separation and preconcentration of uranium (VI) and thorium (IV). J. Hazard. Mater. 152, 79–84.

N.-E. Belkhouche, M.A. Didi / Hydrometallurgy 103 (2010) 60–67 Shemirani, F., Baghdadi, M., Ramezani, M., Jamali, M.R., 2005. Determination of ultra trace amounts of bismuth in biological and water samples by electrothermal atomic absorption spectrometry (ET-AAS) after cloud point extraction. Anal. Chim. Acta 534, 163–169. Star, J., Dahdouh, H., Shlewit, H., Khorfan, S., 2002. Statistical study of factors affecting the co-extraction of uranium and iron in the second cycle of extraction with DEHPA/TOPO in kerosene. Hydrometallurgy 65, 23–30. Taher, M.A., Rezaeipour, E., Afzali, D., 2004. Anodic stripping voltammetric determination of bismuth after solid-phase extraction using amberlite XAD-2 resin modified with 2(5-bromo-2-pyridylazo)-5-diethylaminophenol. Talanta 63, 797–801.

67

Thirupathi, P., Kim, S.S., 2009. Three components synthesis of homoallylic amines catalyzed by bismuth(III) nitrate pentahydrate. Tetrahedron 65 (27), 5168–5173. Yang, J.G., Yang, J.Y., Tang, M.T., Tang, C.B., Liu, W., 2009. The solvent extraction separation of bismuth and molybdenum from a low grade bismuth glance flotation concentrate. Hydrometallurgy 96 (4), 342–348.