Factorial experimental design for the removal of disperse dyes using hydroxyapatite prepared from Moroccan phosphogypsum

Azo dyes are the major known group of synthetic dyes and have given rise to many water and soil environmental problems, most of these azo dyes were used in the textile industry. The aim of this study is the removal of Disperse Blue 79 (DB 79) and Disperse Blue 165 (DB 165) as azo dyes by Hydroxyapatite (HAP). The adsorption experiments were carried out to investigate the factors that influence the dyes uptake by hydroxyapatite, such as the contact time under agitation, adsorbent dosage, initial dye concentration, and pH solutions. To reduce the number of experiments, full factorial experimental design at two levels (2) was used to achieve optimal conditions for the removal of DB 79 and DB 165 from aqueous solutions.


Introduction
Azo dyes are widely used in textile and dyeing industries worldwide.The presence of these dyes at very low quantities in wastewater is highly visible and undesirable.However, it becomes toxic in the water bodies because of its complex molecular structure with multiple aromatic groups 1 and because many of these dyes are made from known carcinogens, such as benzidine and other aromatic compounds.
Adsorption is considered an attractive technique used for the removal of dyes from water.Many adsorbents were used to remove dyes, but hydroxyapatite (HAP) is an important inorganic material in chemistry and biology 11 , their ionic exchange property, adsorption affinity, and their ability to form bonds with organic molecules of different sizes, have given to this material the capability to attract different kinds of dyes.
The present study aimed to determine the optimum conditions for the removal of disperse dyes, DB 79 and DB 165 from aqueous solutions by synthesized hydroxyapatite from phosphogypsum.The factors that influence the dye uptake by the adsorbent were studied and optimized by Nemrodw (2000).

Reagents
All chemicals used in the preparation and the adsorption studies were of analytical grade and used as received without further purification.Two different leather dyes (Disperse Blue 79 and Disperse Blue 165) were purchased from Sigma-Aldrich and carbosynth chemicals; the product number is 12239-34-8 and 41642-51-7 respectively.The chemicals structures of these dyes are shown in Figure 1.

Preparation of HAP
The phosphogypsum was treated using sulfuric acid 67%, the obtained composite was washed by water, by acetone, dried at 65°C and the final anhydrite was sieving at 40 µm 12 .The sieving anhydrite was mixed with phosphoric acid.The pH of the reaction was adjusted tobe 11 by adding Hydroxide sodium.After that, the produced HAP has been removed from the solution by filtration and the resulting powder was washed by water, dried at 80°C and 900°C 13 .

Adsorption experiments
All adsorption experiments were carried out in 150 mL Erlenmeyer flasks by mixing HAP with 100 mL of the dye solution at the desired operating conditions.Flasks were shaken in a stirrer (Velp scientifica) at 20°C.The dye solutions were stirrer along the time, and these were centrifuged at 4000 rpm.The supernatant was separated to measure the dyes concentration using JASCO V-630 spectrophotometer.For this measurement, the wavelengths of maximum absorption were 537 and 540 nm for DB 79 and DB 165 respectively.
The adsorption percentage (%Y) and adsorption capacity values at equilibrium, qe(mg.g - )and time t, qt(mg.g - ) were calculated using the following equations: Where C0 (mg/L) is the initial dye concentration, Ct is the dye concentration at time (t), Ce is the dye concentration at equilibrium, and W (g) is the HAP amount in the solution, V (l) is the volume.

Experimental design and statistical analysis
In order to study the effects of the various parameters, 2 4 factorial experimentations were carried out, in two levels (low (-) and high (+)).The effects were designed as in Table 1 which shows the values of the factors selected in this study.
Table 1.Independent variables and levels used in this study.This factorial experimentation results in sixteen tests with all possible combinations of X1, X2, X3 and X4 14 .Removal of DB 79 (%YDB79) and Removal of DB 165 (%YDB 165) were calculated for each of these tests as shown in Table 2.

Levels
The polynomial equation based on the first-order model with four factors (X1, X2, X3 and X4) and all possible interactions was given in the form of the following expression: Where b0 is the average value of the result; b1, b2, b3 and b4 are the linear coefficients; and b12, b13, b14, b23, b24, b34, b123, b124, b134, b234 and b1234 represent the interactions coefficients.A total of sixteen experiments were analyzed using NEMRODW version 2000 15 .Fig. 2.a Shows XRD spectra of HAP, the formation of hydroxyapatite was indicated by characteristic peaks occurring at 2θ=25.9°, 31.9°,32.9°, and 34.1°, all diffraction peaks were indexed to pure hexagonal structural HAP (space group P63/m) according to JCPDS card no.09-043211.Fig. 2.b shows FTIR spectra of HAP, PO4 3− bands were registered from 500 to 600 cm −1 , the bands assigned to the stretching modes of hydroxyl groups in hydroxyapatite were detected at 3560 cm −1 , the stretching bands of CO3 2-have been reported at 1360 cm −1 12 .

Study of the influence of factors on adsorption of DB 79 and DB 165 by HAP
The data obtained from the removal of DB 79 and DB 165by HAP (Table 2) suggested that assays 12 and 5 presented higher and lower adsorption removal of DB 79; however assays 9, 11, 15 and 6 presented higher and lower adsorption removal of DB 165.

Data analysis
Removal of DB 79 By combination between the different factors, 2 4 experiences were listed in Table 2.The results obtained (Fig. 3) show that X1 (concentration of DB 79) and X3 (size) have a negative effect on the Y1 (% Removal of DB 79) response: decreasing the concentration of DB 79 and size of HAP (100-20 mg/L and 200-125µm, respectively) results in increase of the % removal of DB 79.The adsorbent concentration of HAP effect is positive; the % removal of DB 79 increases by increasing the adsorbent concentration, and this is due to the surface area available by more adsorbent particles 16

Experimental design techniques-Doehlert matrix
The Doehlert design is a response surface methodology based on multivariate optimization of variables 17 .The statistical design of adsorption of DB 79 and DB 165 by HAP was used to determine the optimum condition for the removal efficiency for both dyes by HAP, which was done using the Doehlert design 18 .The aim of this step was to determine the best conditions for the adsorption of DB 79 and DB 165 by HAP, the adsorbent concentration was fixed at the high value (5g/L) per 100 ml of solution and size granulometry was fixed at 125 μm.This study was performed to focus on the predominant factors: the concentration of dyes and contact time, in the domain described in Table 3.The matrix and corresponding code were given in Table 4.
According to the methodology of Doehlert matrix, a second order polynomial response equation was used to correlate the response and independent variables: Where Y was the percentage removal of dye, b0, bi, bii, bij were the constant coefficients and Xi,j independent variables 19 .Statistical analyses were performed using nemrodw 2000.Two responses (the yield (Y1), the percentage removal of DB 79 and (Y2), the percentage removal of DB 165 were studied and simultaneously optimized as described in Table 3.

Variables optimization Removal of DB 79
The removal of DB 79 using HAP has been followed using UV-vis spectrophotometer.The analysis of this response (Table 4 and Fig. 7) showed that the percentage of DB 79 (Y1) varies between 73 and 97%.The DB 79 adsorption (Y1) can be described as follows: The correlation coefficient calculated by the model is adequate (i.e.R 2 = 0.981for DB79, and R 2 A = 0.965 for DB165 respectively).The contact time is the principal factor influencing in adsorption of DB 79 by HAP; when the contact time increases, the DB 79 removal increases (Fig. 7).To maximize the adsorption of DB 79 by HAP, the contact time between the adsorbent and the adsorbate must increase.The best performances were obtained, close to the central point, X1 (-1) and X2 (+1).In particular, low removal efficiencies were obtained for high concentration of DB 79 (X1 = +1) and low contact time (X2 =−1).

Comparison of contact time for adsorption of various dyes using hydroxyapatite
The optimum contact time for the adsorption of various dyes using hydroxyapatite prepared from different synthesis method is presented in Table 5.A comparison between this work and other reported data from the literature shows that synthesis of HAP can be done by various techniques using different reactants and production routes that can modify its physicochemical properties, morphology, chemical composition homogeneity, particle sizes, and degree of crystallinity.Hence, different contact time of adsorption and different removal uptake can be registered, in our work, HAP was prepared using a low-cost method with satisfactory results on removing disperse dyes.
. The contact time is positive, the % removal of DB 79 increases by increasing contact time.The interactions are represented graphically in Fig. 4 a, 4 b and 4c.The % removal of DB 79 increases by increasing the adsorbent concentration of HAP, contact time and by decreasing the initial DB 79 concentration and with 125 μm in size of HAP.

Figure 3 .
Figure 3. Values of the linear and quadratic coefficients for % removal of DB 79

Figure 4 .
Figure 4. Interaction diagram for response Y1: Yield.(a) adsorbent concentration/Concentration DB 79, (b) Size/ Concentration DB 79 and (c) Time/ Concentration DB 79.Removal of DB 165The analysis of Fig.5shows that the X2 (adsorbent concentration), X3 (size) and X4 (time) have a positive effect of DB 165 removal, the % removal of DB 165 increased by increasing adsorbent concentration, size of HAP and contact time, however

Figure 5 .
Figure 5. Values of the linear and quadratic coefficients for % removal of DB 165

Figure 7 .
Figure 7. Variation of total yield (Y1, %), of time as a function of concentration of DB 79: (a) 2D, (b) 3D Removal of DB 165 Fig. 8 displays the dependence of the % removal of DB 165 as a function of contact time and DB 165 concentration.The removal % of DB 165 was only slightly affected by DB 165 concentration with a

Table 2 .
Experimental design in term of factors and results of responses.

Table 3 .
Independent variables and levels used in Doehlert's matrix.

Table 5 .
The optimum contact time to reach equilibrium for the adsorption of various dyes using hydroxyapatite as an adsorbent.The adsorption of DB 79 and DB 165 using hydroxyapatite prepared from phosphogypsum has been studied.The full factorial design applied allowed the evaluation of the most important factors for the removal of DB 79 and DB165.The main conclusions that can be drawn from this study are given below:Analysis of different effects of factors (contact time, size, initial dye concentration and adsorbent concentration) shows that the most influential factors, on the yield of removal of DB 79 and DB 165, are contact time and initial dye concentration. Response Surface Methodology was used to determine the acceptable compromise zone for responses of adsorption of DB 79 and DB 165 by hydroxyapatite.The optimal conditions were identified the time at 23 min and 60 initial DB 79 and DB 165 concentrations.At these conditions, the removal percentage of DB 79 and DB 165 were 94 and 98 %, respective.