Choice of factors

An adequate choice of parameter variation domains is an essential condition for establishing a precise model that perfectly describes the studied process [8]. Based on a bibliographic analysis [9][10], we developed a complete factorial plan with a number of factors equal to four namely drying temperature (X1), biosorbent granulometry (X2), stirring time (X3) and concentration (X4). Each factor takes two levels namely the low level (−1) and the high level (+1). The various factors and their areas are summarized in Table 1.

Table 1. Factors and their different areas of variation.

Different stages of biosorbent preparation

The biosorbent used has been pre-treated to remove sand and coarse elements. Subsequently, the preparation required several stages of treatment:

Drying

Rubber hulls (Figure 1) are dried in an oven at two different temperatures for 24 hours at 75˚C and 120˚C. The rubber hulls used in this study were collected in the Daloa area, a city in Côte d’Ivoire.

Grinding and sieving

Rubber hulls are dried and crushed (Figure 2) using a crusher and then sieved with sieves of diameters 1 mm; 0.5 mm and 0.2 mm to obtain a grain size between 0.2 and 0.5 mm then between 0.5 and 1 mm (Figure 3).

Figure 1. Rubber hulls.

Figure 2. Rubber hulls after drying (a) and grinding (b).

Figure 3. Rubber hulls after sieving. (a) Rubber hulls after sieving with a diameter between 0.5 and 1 mm; (b) Rubber hulls after sieving with a diameter between 0.2 and 0.5 mm.

Washing

After grinding, the biosorbent was washed at room temperature. This consists of removing all impurities and soluble substances from the samples (pigment). Washing is carried out first with tap water until the washing water is colorless. The samples are then bleached with a sodium hypochlorite (bleach) solution of concentrations 0.107 mol∙L⁻¹ and 0.214 mol∙L⁻¹, and rinsed with distilled water. At the end of washing the biosorbents are dried in an oven at 75˚C and 120˚C for 24 hours.

Methylene blue index

Methylene blue index (mg/g) is an indicator of the biosorbent’s ability to adsorb medium and large organic molecules. It characterizes the mesopores of the crushed rubber hull. For the determination of this index, the method of the European Centre of Chemical Industry Federations from the work of Mamane et al. [11] was used. In a 250 mL erlenmeyer, 0.1 g of crushed rubber shell (biosorbent) and 100 mL of 1944 × 10⁻⁵ M methylene blue solution is introduced.

The mixture is stirred for 20 min then filtered. The residual methylene blue concentration is determined using a UV-Visible spectrophotometer at a wavelength of 620 nm. Thus, the methylene blue index (Q_MB) is given by the following relation:

With:

V: Volume of methylene blue solution in (mL)

M: Methylene blue molar mass (g/mol)

Ci: Initial methylene blue concentration (mol/L)

Cr: Residual methylene blue concentration in (mol/L)

m_b: Mass of biosorbent (g).

Q_MB: Methylene blue index (mg/g)

Iodine index

The purpose of the iodine index is to determine the ability of the crushed rubber shell to adsorb small molecules. It characterizes micropores accessible to small particles. This test was performed according to the AWWA B 600-78 standard from the work of Maazou et al. [12]. In a 100 mL beaker, 0.05 g of crushed rubber shell (biosorbent) is introduced. Add 20 mL of 0.1 N iodine solution to the pipette and stir for 5 min before filtering. A volume of 10 mL of the filtrate is taken and put into an erlenmeyer. From the burette, a solution of sodium thiosulphate with a concentration of 0.1 N is added to the erlenmeyer containing the filtrate until the total discoloration of the solution. Starch is used as a coloured indicator. Iodine index (Q_II) is given by the following relationship:

With:

C₀: Initial concentration of iodine solution (mol/L)

C_tĭio: Concentration of sodium thiosulfate solution (mol/L)

V_tĭio: Volume of thiosulfate poured at equivalence (mL)

V_I2: Iodine volume (mL)

M_I2: Molar mass of diiodine (g/mol)

Vads: Sample filtrate volume (mL)

m_b: Crushed rubber shell mass (g)

Q_Ii: Iodine index (mg/g)

Results are processed using multiple linear regression and analysis of variance. It states that the response is a linear function of all factors. Thus, to determine this function is to find the coefficients of the following polynomial equation:

With:

a_i: the effect of factor Xi

a_ij: the effect of two-fold interactions

a_ijk: the effect of three-fold interactions

a_ijkl: the effect of four-fold interactions

X_iX_j: the interaction between X_i and X_j

The determination of the experimental domain, experimental designs, estimates and statistics of the polynomial coefficients is carried out using the MINITAB 19 software.

The experimental results of the methylene blue index and the iodine index obtained according to the complete factor design are reported in Table 2.

Methylene blue shows between 2.80 and 4.88 mg/g, while iodine shows between 203.05 and 263.96 mg/g.

Table 2. Experiment matrix and experimental data.

Values for iodine are higher than those for methylene blue. This shows that biosorbents from rubber hulls have more capacity to adsorb small molecules than medium molecules. The maximum ID value is obtained for a biosorbent with a diameter between 0.2 and 0.5 mm, regardless of its drying temperature. The results of the iodine index obtained are similar to those of Benamraoui [13] recorded during his work on biosorbents from apricot and olive kernels.

In addition, our work generally shows that the highest methylene blue index values are obtained with small diameter biosorbents (0.2 and 0.5 mm). These results reveal that particle size determines the adsorption rate; the smaller the grains, the faster the transfer to the center [14].

The estimated coded coefficients are given in Table 3 and Table4.

Table 3. Coding coefficients for methylene blue index (IBM).

From Table 3, it is noted that the p-values obtained are less than 0.05 (p < 0.05) for drying temperature (X1) and grain size (X2). This shows that these two parameters have a very significant effect on the adsorption capacities of the biosorbent. Also, the interaction between the particle size and the concentration of sodium hypochloride X2*X4 and the interaction between the stirring time and the mass of sodium hypochloride X3*X4 have p values less than 0.05. This suggests that these interactions greatly influence the adsorption capacities of the medium molecules of the biosorbent from rubber shells.

In addition, the coded regression equation linking parameter levels to IBM is given by:

IBM = 3.9083 + 0.1997X₁ − 0.5078X₂ + 0.0807X₃ + 0.0685X₄− 0.0990X₁*X₂ − 0.0130X₁*X₃+ 0.0406X₁*X₄ − 0.0270X₂*X₃ + 0.2113 X₂*X₄ + 0.1256X₃*X₄

Table 4. Coded iodine index coefficients.

The results obtained for the iodine index (ID) response coefficients show that parameters such as temperature (X1), grain size (X2), stirring time (X3) and sodium hypochloride concentration (X4) have p values less than 0.05 (Table 4). Indeed, these parameters have an important influence on the adsorption capacities of a biosorbent from the rubber husks to adsorb small molecules. Also, the interaction between the particle size and the mass of sodium hypochloride (X2*X4), then the interaction between the stirring time and the mass of sodium hypochloride (X3*X4) and then the interaction between the temperature and the stirring time have p values less than 0.05. These different results show that these interactions strongly influence the adsorption capacities of the small molecules of the biosorbent from the rubber shells.

Equation de régression codées reliant les niveaux paramètre avec indice d’iode est:

ID = 232.60 + 8.15X₁ − 6.23X₂ + 5.43X₃ + 5.80X₄ − 1.16X₁*X₂ − 4.24X₁*X₃ − 3.00 X₁*X₄− 5.13X₂*X3 + 1.13X₂*X₄ + 5.24X₃*X₄

It appears from the analyses of these two tables above that the adsorption capacities of the small molecules are larger than those of the average molecules with the biosorbent from the rubber shells.

Analysede la variance

The results of Table 5 show that particle size (X2), is the most important factor in the optimization of the biosorbent on the methylene blue index (IBM). Indeed, the contribution of grain size to the elaboration of the answer is 59.41%.

All the studied parameters influence the optimization of the biosorbent on the iodine index (ID). However, none of these parameters contribute significantly to the development of the response (Table 6).

Table 5. Variance analysis results for methylene blue index (IBM).

Table 6. Variance analysis results for iodine index (ID).

Figure 4. Pareto diagram of normalized methylene blue index (IBM) effects.

Direct effects of parameters on methylene blue index

The Pareto diagram (Figure 4) shows that drying temperature appears to be the most influential factor in the optimization of biosorbent from rubber hulls. Then comes the particle size which also contributes to this optimization of the biosorbent. This observation was made by Mamane [11] and Aboua [15] during their work on the optimization of the conditions of active charcoal production. According to this work, small particles have a large specific surface area and therefore lead to better adsorption and shortening of the equilibrium time. The results of the diagram show that temperature and particle size strongly influence the biosorbent’s ability to adsorb medium molecules. Also, the effect of the interactions between the granulometry and the sodium hypochloride mass, and between the stirring time and the sodium hypochloride mass is significant on the optimization of the biosorbent to the adsorption of medium molecules.

Direct effects of parameters on iodine index

The Pareto diagram (Figure 5) shows the existence of several influencing factors. Indeed, all the studied parameters strongly influence the biosorbent’s ability to adsorb small molecules. In addition, certain interactions between the different parameters also contribute to the optimization of the biosorbent in the adsorption of small molecules.

The diagram of the main effects on IBM (Figure 6) confirms that drying

Figure 5. Pareto diagram of standardized iodine index (ID) effects.

Figure 6. Main effects of methylene blue index graph.

temperature and particle size strongly influence the adsorption capabilities of the biosorbent to adsorb medium molecules. In the diagram of the main effects on ID (Figure 7), we also observe that all parameters have a positive influence on the adsorption capacities of small molecules.

Effect of factor interactions

There are interactions between granulometry and sodium hypochloride mass that influence the adsorption of medium molecules (Figure 8). This interaction would modify the functional groups on the surface of our biosorbent. In addition, interactions are observed between the majority of the parameters that contribute to modify the adsorption capacities of the biosorbent for small molecules (Figure 9).

Outline diagrams

Figure 7. Iodine index major effects graph.

Figure 8. Methylene blue index interactions diagram.

Figure 10 and Figure 11 show the drying temperature contour areas (X1) and the grain size (X2) that maximize the IBM and ID responses respectively by keeping the stirring time (X3) and the sodium hypochloride mass (X4) constant. Indeed, the contour areas that optimize these responses are close to 120˚C and a grain size between 0.2 and 0.5 mm. This suggests that a high temperature has a significant influence on the microporosity of absorbent thus improving their adsorption capacity. According to Park et al. [16], a higher temperature generally improves the response but can damage the physical structure of the biosorbent.

Figure 9. Iodine index interactions diagram.

Figure 10.Methylene blue index contour diagram.

Multi-response optimization

Figure 12 shows the optimization results of biosorbent preparation on IBM and ID. We find that the answers have the same optimization values for the different parameters which are X1 = 1; X2 = −1; X3 = 1; X4 = 1. These results are acceptable because it is desirable to remain in the optimal conditions comparable

Figure 11.Iodine index contour diagram.

Figure 12. Multi-response optimization graph.

to those recorded in order to obtain an optimal value and optimal condition for biosorbent preparation [17]-[19]. It can also be noted that the desirability value (d) of IBM and the ID is 1. However, the closer the value of d is to 1, the more efficient the optimization is (Rani, 2015) [4].