RR120 with 99 % of purity (molar weight 1469.98 g/mol, CAS-No 61951-82-4 and linear formula C₄₄H₂₄Cl₂N14Na₆O₂₀S₆) was obtained from Sigma-Aldrich and was used as adsorbate without further purification. Figure 1 shows that it contains two azo groups as chromophores, two chlorotriazine groups as reagents, six sulfonic acid groups and two phenolic OH groups (Paul et al. 2011).

Fig. 1

The adsorbent solids were prepared by the continuous co-preparation method previously reported by Perez-Lopez et al. (2006). Two aqueous solutions were prepared: solution A containing nitrates of the metals Mg(NO₃)₂.6H₂O and Al(NO₃)₃.9H₂O for HDL-MgAl; and of Mg(NO₃)₂.6H₂O and Fe(NO₃)₃.9H₂O for HDLMgFe. Solution B was composed by a mixture of NaOH and Na₂CO₃. The atomic ratio between bivalent and trivalent metal cations (M²⁺/M³⁺) was maintained at 3. For the synthesis of HCMg, solutions A and B were composed of Mg(NO₃)₂.6H₂O and Na₂CO₃, respectively. These two solutions were mixed simultaneously in a glass continuous stirred-tank reactor (CSTR), with constant temperature and pH. The precipitate was maintained under agitation for 1 h at 50 ºC, and then vacuum filtered, washed with deionized water and dried in an oven at 80 ºC overnight.

The solids were characterized by X-ray diffraction (XRD), N₂ adsorption/desorption (BET) and temperature-programmed desorption (TPD-CO₂).

The X-ray diffraction patterns were collected through the powder method with a Bruker D2-phaser diffractometer using Cu-Kα radiation (λ = 1.5406 Å) at 30 kV and 10 mA, 2 q between 5 and 70º, and step of 0.02º s^-1.

The surface area measurements were performed on a NOVA 1000e equipment (Quantachrome Instruments). Nitrogen was used as inert gas, and a pretreatment with vacuum for 1 h at a temperature of 300 ºC was performed.

TPD-CO₂ profiles were performed in a multipurpose system using 100 mg of the sample. Initially, the samples were degassed at 100 ºC with a helium flow, for 30 min. Then, samples were saturated with 30 mL/min of CO₂. After adsorption, the samples were purged with a pure helium stream to remove species that were not adsorbed. Subsequently, the heating was initiated at 10 ºC/min with helium flow of 30 mL/min. The desorption curves were recorded with a thermal conductivity detector (TCD).

The adsorption tests were performed in batch mode using volumes of 100 mL of RR120 solutions. In each test, the adsorbent was added to the solution and the process parameters to obtain the best experimental conditions were evaluated. All experiments were performed in duplicate. Considering that the prepared adsorbents have alkaline properties, and the dye used as a pollutant is acidic in nature so the intensity of its color is affected by sudden changes in pH, the experiments were carried out at the natural pH of the solution containing the solids, which was approximately 8.

The best experimental conditions were obtained by adsorption experiments using 100 mL of an RR120 solution (30 mg/L), and HDL-MgAl, HDL-MgFe and HC-Mg as adsorbents. The effect of adsorbent dosage was investigated considering different dosages of the three solids (0.05, 0.1, 0.15, 0.2, 0.25, 0.3 and 0.35 g). For these tests, the adsorption contact time was 30 min. Contact time experiments (5, 10, 20, 40, 60, 90 and 120 min) were performed applying the best adsorbent concentration obtained. All experiments were performed in duplicate and the solutions were centrifuged in a CIENTEC equipment, model CT5000R, at 6000 rpm for 10 min and analyzed in a spectrophotometer at a wavelength of 535 nm. The amount of dye adsorbed was calculated by equation 1.

q t = ( C 0 - C t )   x   V m (1)

where C ₀ (mg/L) was the initial concentration of dye, C _t (mg/L) was the concentration of dye at time t, q _t (mg/g) was the adsorbed amount at time t, V (L) was the volume of phosphate solutions, and m (g) was the dosage of the adsorbent solids.

The kinetic data were analyzed using pseudo-first order and pseudo-second order models. Isotherms were constructed by conducting experiments varying the concentration of the dye solution (10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400 and 500 mg/L). Solids dosage and adsorption time were used according to the last item. Three different isotherm models were tested to fit the equilibrium data: Langmuir, Freundlich and Redlich-Peterson.

For the parameters estimation of each model, equation 2 was used to minimize errors:

E r r o r = q e c a l c - q e e x p q e e x p 2   (2)

where q _e ^calc and q _e ^exp represent the amount adsorbed in equilibrium calculated by the models and the experimental values, respectively.

The XRD patterns of the adsorbents can be observed in figure 2. The diffractograms of HDL-MgAl and HDL-MgFe present peaks corresponding to the structure of the hydrotalcites, which show three main peaks at approximately 11, 22 and 35º. These reflections correspond to diffraction by planes (003), (006) and (009) and represent the separations between layers (Vicente 2002). These materials present common features such as the presence of intense and sharp peaks for low angle values and less intense peaks for larger angle values that are generally asymmetric (Cavani et al. 1991).

Fig. 2

The diffractogram for HC-Mg shows peaks corresponding to hydromagnesite or magnesium hydroxycarbonate, similar to those reported in the literature by Beall et al. (2013). The crystallinity of the structure is evidenced by the two main reflections at approximately 15 and 30º.

The lamellar double hydroxides synthesized presented good crystallization and a structure between layers similar to those reported in the literature (Tsyganok and Sayari 2006, Guo et al. 2013, Shan et al. 2015). Comparing the HDL-MgAl and HDLMgFe diffractograms, the sample with Al forms a hydrotalcite with higher crystallinity and more intense reflections than THE sample with Fe.

Table I presents the results of the specific surface area (BET) and the number of basic sites obtained in the TPD-CO₂ measurements. Hydrotalcite-type samples show high surface area values in comparison with HC-Mg. The main reason for this difference is due to the structure of lamellar double hydroxides having a large surface with spaces between layers (Vaccari 1998). The iron-containing hydrotalcite (HDL-MgFe) has a larger surface area than the aluminum-containing hydrotalcite (HDL-MgAl), which is in accordance with the lower crystallinity of HDL-MgF revealed in the XRD pattern (Fig. 2). This fact is important because the ability to solute removal is also related to the available surface area in the adsorbent and to the size of the adsorbate molecule. The pore diameter is also important because for smaller pores the adsorption of larger molecules would present difficulties of access to the internal pores of the solid.

TABLE I

Samples	Surface area BET (m2/g)	Maximum temperature of the basic peak (ºC)	Total number of basic sites (mmol/g)
Hydrotalcite of MgAl	75	335	0.3
Hydrotalcite of MgFe	119	385	0.2
Hydroxycarbonate of Mg	67	303	0.7

The number of basic sites on the surface of the samples was obtained by the integration of desorption peaks of TPD-CO₂. The strength of the sites can be determined by the temperature at which the desorption peak occurs. Note that samples reveal maximum desorption temperatures (Tmax) above 280 ºC, which corresponds to medium (240-303 ºC) and strong (326-396 ºC) sites (Pavel et al. 2012)Fe, Co, Ni, Cu and Zn.

Table I shows the HC-Mg sample exhibited the highest total number of basic sites. When Al or Fe was added to Mg, there was a decrease in total sites for both samples (Carvalho et al. 2012). However, Tmax increased with the addition of the trivalent cation in the Mg structure, that is, the HDL-MgAl and HDL-MgFe samples showed an increase in the strength of basic sites.

Based on desorption temperatures in TPD, the comparison between the hydrotalcites showed that HDL-MgFe provides strong basic sites than HDL-MgAl. However, samples containing Fe exhibit lower total number of basic sites than samples with Al.

Also, it is important to highlight that the effect of pH is significant in the adsorption stud, although the intensity of the RR120 solution color was affected by sudden changes in pH. The consideration of the alkaline properties through pH_PZC is an important parameter for the solid adsorption capacity. The pH_PZC value is 11 for HDL-MgAl and HCMg, and 8 for HDL-MgFe solid. In this case, as RR120 is an acid dye, the hydroxyl group on the adsorbents surface can interact with the dye protons, providing a favorable electrostatic interaction in the removal process.

Adsorption experiments of RR120 dye in aqueous phase were performed at different adsorbent dosages. Figure 3 shows the effect of the adsorbent dosage on the removal of RR120 using the HDL-MgAl, HDL-MgFe and HCMg solids.

Fig. 3

Figure 3 shows the concentration of the adsorbed dye increases with the increasing adsorbent dosage. Removal depends on the initial concentration of dye, which was approximately 30 mg/L. It may be evident in figure 3 that after 1 g/L for HDL-MgAl and HC-Mg, and after 2 g/L of HDL-MgFe, equilibrium was attained, which indicates the saturation of the solid surface. However, at a low adsorbent dose (0.5 g/L), HDL-MgAl provided a greater removal of dye (78.8 %). The maximum adsorbent value was obtained by HDL-MgFe, with a removal percentage of 96.9 %. This fact may be related to the larger surface area of the HDL-MgFe hydrotalcite, which results in greater total dye removal capacity.

The maximum removal was reached at adsorbent dose of 3.5 g/L with removal values of 93, 96, and 95 % for HDL-MgAl, HDL-MgFe and HC-Mg, respectively. Therefore, in terms of costs and considering the equilibrium evidenced in figure 3, it is clear that with an adsorbent dose of 2.5 g/L for HDL-MgAl and HC-Mg, and 2 g/L for HDL-MgFe, a dye removal near 90 % was reached. These values are very close to the maximum removal achieved in the tests.

Figure 4 shows the removal of RR120 dye at different adsorption times. The adsorbent dosage for solids was 2.5 g/L for HDL-MgAl and HCMg, and 2.0 g/L for HDL-MgFe.

Fig 4

The evaluated solids show very fast kinetics for the process, since after 5 min more than 60 % of the dye was adsorbed with HDL-MgFe, and above 80 % with HDL-MgAl and HC-Mg. This result is important, since a rapid removal of adsorbate and achieving equilibrium in a short time indicate that the adsorbents used are efficient. The lowest percentage removed with HDL-MgFe in this time interval may be related to the lower dose of adsorbent used.

It can also be observed that for the three solids tested, the kinetics of dye removal is faster during the first 20 min, but after a period of 60 min dye removal occurs slowly and the variation of concentrations is not relevant over time. This happens because initially all the active sites are free at the surface of the adsorbent solids, which results in a fast adsorption. When occupied by the sorbate, free sites decrease, leading to saturation of the solids. Therefore, it was determined that the most suitable time for the removal of RR120 is 60 min for the three solids studied.

Knowing the effect of contact time on the RR120 adsorption experiments made it possible to study the kinetics of the process. The pseudo-first-order model (Lagergren 1898) describes the adsorption rate based on the adsorption capacity. It is expressed by the following equation:

q t = q 1 ( 1 - e x p ⁡ ( - k 1 t ) (3)

The pseudo-second-order equation is expressed as follows (Ho and McKay 1999):

q e = q m k L C e 1 + k L C e (4)

The calculated parameters of the kinetic models are shown in table II, where it can be seen that the values obtained for RR120 adsorption were 27.07, 14.08, and 34.07 mg/g for HDL-MgAl, HDL-MgFe, and HC-Mg, respectively. The kinetics of dye adsorption followed the pseudo-second-order model because its determination coefficients were higher (R² = 0.9954, 0.9966, and 0.9979 for HDL-MgAl, HDL-MgFe and HC-Mg, respectively). This model suggests that chemisorption is a rate limiting step of adsorption. Electrostatic attraction, anion exchange and chemical bonding may be involved in the dye adsorption process (Shan et al. 2015). The same model was found for the adsorption of RR120 in other studies (Tabak et al. 2010, Demarchi et al. 2013, Shan et al. 2015).

TABLE II

Adsorbent solid	Pseudo-first order model	Pseudo-second order model
HDL-MgAl q exp = 27.84 mg/g	q1 (mg/g) = 27.07 k1 (min-1) = 2.21 R2 = 0.993	q2 (mg/g) = 27.10 k2 (min-1) = 45566.38 R2 = 0.995
HDL-MgFe qexp= 14.66 mg/g	q1 (mg/g) =14.08 k1 (min-1) = 0.18 R2 = 0.978	q2 (mg/g) = 15.00 k2 (min-1) = 980.00 R2 = 0.99
HC-Mg qexp= 34.92 mg/g	q1 (mg/g) = 34.07 k1 (min-1) = 0.38 R2 = 0,9894	q2 (mg/g) = 35.13 k2 (min-1) = 41792.03 R2 = 0,9979

*q _exp is the experimental capacity of adsorption, q ₁ and q ₂ are the pseudo-first rate and pseudo-second order constants, respectively, q _exp is the amount of solute adsorbed at equilibrium, and k ₁ and k ₂ are the velocity constants

In order to establish the most appropriate correlation for the equilibrium curves and to estimate the parameters for an adsorption system it is important to establish the most appropriate correlation for the equilibrium curves. The adsorption capacity at equilibrium can be obtained by measuring the adsorption isotherm of the adsorbent (Cheung et al. 2009).

Therefore, to establish the most appropriate correlation for the equilibrium curves and to estimate the parameters of the isotherms, the Langmuir equation (4) (Langmuir 1918), the Freundlich equation (5) (Freundlich 1906), and the Redlich-Peterson equation (6) (Wu et al. 2010) were used.

q e =   K F C e 1 / n (5)

q e = K R C e 1 + a R C e N (6)

where q _e (mg/g) is the equilibrium adsorption amount at equilibrium concentration of Ce (mg/L); qm is the maximum capacity of the adsorbent (mg/g); K _L is the Langmuir adsorption constant (L/mg); K _F (mg¹⁻ⁿ/Ln/g⁻¹) and n are the Freundlich affinity coefficient and linearity, respectively; K _R is the Redlich-Peterson adsorption constant (L/mg), aR is a Redlich-Peterson isotherm constant (L/mg), and N is a Redlich-Peterson isotherm exponent.

In order to predict whether the adsorption of RR120 in aqueous solution was efficient or not, the shape of the isotherm, the statistical parameters and the values of the constants for each non-linearized model were taken into consideration.

Figure 5 shows the experimental adsorption isotherms for the removal of RR120 using HC-Mg as adsorbent and compares them with the Langmuir, Freundlich and Redlich-Peterson models.

Fig. 5

Table III shows that the Langmuir and Redlich-Peterson models best described the adsorption equilibrium data of RR120 because they presented R² values close to 1.

TABLE III

Freundlich	Langmuir	Redlich-Peterson
k F = .16.48	q max = 45.94	q m = 43.88
n = 5.23	k L = 0.27	k R = 0.29
		n = 0.99
R2 = 0.87	R2 =0.98	R2 = 0.98
Error = 1.27	Error = 0.15	Error = 0.15

*k _F , k _L and k _R are Freundlich, Langmuir and Redlich-Peterson constants, respectively; n is an isotherm exponent; q _m and q _max are the maxima capacities of the adsorbent (mg/g)

Figure 6 shows the experimental adsorption isotherms compared with the Langmuir, Freundlich and Redlich-Peterson models for the removal of RR120 using HDL-MgAl.

Fig. 6

Table IV reports the isotherm parameters evaluated for HDL-MgAl in the removal of RR120. As it can be seen, the Langmuir and Redlich-Peterson models best described the adsorption equilibrium data for the removal of RR120 with this solid, since an R² value close to 1 and low errors were obtained.

TABLE IV

Freundlich	Langmuir	Redlich-Peterson
k F = 28.38	q max = 108.83	q m = 120.04
n = 3.69	k L = 0.28	k R = 0.24
-	-	n = 1.021
R2 = 0.91	R2 = 0.98	R2 = 0.98
Error = 3.47	Error = 0.47	Error = 0.45

*k _F , k _L and k _R are Freundlich, Langmuir and Redlich-Peterson constants, respectively; n is an isotherm exponent; q _m and q _max are the maxima capacities of the adsorbent (mg/g)

Similarly, figure 7 shows the experimental adsorption isotherms for the removal of RR120 and compares them with the Langmuir, Freundlich and Redlich-Peterson models in the case of HDL-MgFe.

Fig. 7

Table V shows that the Redlich-Peterson model best described the equilibrium data of the RR120 with HDL-MgFe, with an R² value close to 1 and a minor error.

TABLE V

Freundlich	Langmuir	Redlich-Peterson
k F = 20.02	q max = 82.59	q m = 48.87
n = 4.48	k L = 0.09	k R = 0.22
		N =0.92
R2 = 0.94	R2 = 0.98	R2 = 0.99
Error = 1.04	Error = 0.34	Error = 0.11

*k _F , k _L and k _R are Freundlich, Langmuir and Redlich-Peterson constants, respectively; n is an isotherm exponent, q _m and q _max are the maxima capacities of the adsorbent (mg/g)

Figures 5, 6 and 7 show the adsorption isotherm models considered for the three solids used in the removal of RR120. The models that have a better correlation with the experimental data are Langmuir and RedlichPeterson, since they adjust the experimental results in a satisfactory way unlike the other models studied, which do not adjust to the data neither describe the thermodynamic behavior of the adsorption process.

It can be noticed that the adsorption capacity increases with the equilibrium concentration of the dye in the solution due to the progressive saturation of the monolayer for the studied solids. The Langmuir and Redlich-Peterson models were present in the three systems, although there was not a significant difference between them. These models indicate that the processes occurred at homogeneous and specific sites.

The results of tables III, IV and V show the higher q _max values for HDL-MgAl, although it is not the solid with the highest surface area. This fact can be explained by the greater number and strength of basic sites of that favors the adsorption of dye with acidic characteristics. Although HC-Mg has the highest number of basic sites, these present lower basic strength, as is observed in table I. >On the other hand, although the basic sites of HDL-MgFe showed higher strength, it has a lower number of sites than HDL-MgAl, indicating a compromise between strength and number of sites. The maximum adsorption capacity is not only the surface influence of the adsorbent solids but also involves the dye chemistry and the acid-base properties of the solids. It should be noted that the experimental conditions used in the adsorption process influence the performance of the adsorbent. In addition, the retention of the dyes by the materials involves several attractive forces, such as ionic interactions, Van de Waals forces, hydrogen bonds and covalent bonds. Depending on the type of the dye, one or more forces act in the adsorption process.

The results of the models were compared with those obtained by other authors using the same dye and different sorbent solids. As can be seen in table VI, the models obtained correspond to the Langmuir and Redlich-Peterson models, as well as the present study.

TABLE VI

Solid adsorbents	Isotherm model	Maximum capacity of the adsorbent (mg/g)	Reference
Hydrotalcite of MgAl	Langmuir Redlich-Peterson	q max = 108.8 q m = 120.0	This study
Hydrotalcite of MgFe	Langmuir	q max = 82.6	This study
	Redlich-Peterson	qm = 48.9
Hydroxycarbonate of Mg	Langmuir Redlich-Peterson	q max = 45.9 q m = 43.9	This study
Hydrotalcite of MgAlCO3	Langmuir	59.5	(Shan et al. 2015)
cetylpyridinium-bentonite	Langmuir	81.9	(Tabak et al. 2010)
chitosan-Fe(III)-crosslinked	Langmuir-Freundlich	433.8	(Demarchi et al. 2013)
Chitosan/modified montmorillonite	Langmuir	5.6	(Kittinaovarat et al. 2010)

As can be seen in the table VI, all models evaluated for the removal of RR120 with different sorbent solids describe a Langmuir equilibrium, where the results obtained by the present study for HDLMgAl (q _m and q_max values of 120.04 and 108.83 mg/g, respectively) are much higher than the rest of the values compared in the same table. The value obtained in the present work is lower than the results obtained by Demarchi et al. (2013)) (433 mg/g), but the solid used by these authors had highly different characteristics. However, the values of q _m and q _max obtained in the present work were superior to those reported by Shan et al. (2015) with HDL-MgAlCO₃, which has similar properties to the HDL-MgAl used here, showing that the basic solids tested in this study are a good alternative for the removal acid dyes such as RR120.