Vermicompost was obtained from Suzan Humus (Suzano-SP, Brazil), congo red dye (CR) was purchased from Vetec Company (Duque de Caxias-RJ, Brazil), while Indigo Blue (IB) was purchased from Tupy Industry (São Paulo-SP, Brazil). Ultrapure water used to prepare all solutions was obtained from a water purification unit Quimis (Q241-22 model, Diadema-SP, Brazil). Several laboratory glasswares were used, including beckers, erlemeyers, kitassato, and volumetric flasks. Others equipments were used such as: laboratory oven (Quimis Q-317 B model, Brazil), particle size sieves (Granutest, Brazil), Analytical balance (Shimadzu AY 220 model, Japan), sputter coater (Shimadzu, IC-50 Ion Coater model, Japan), scanning electron microscope (Shimadzu, SSX 550 model, Japan), infrared spectrophotometer (Cary 630 model, Agilent Technologies, Santa Clara, California, USA), pHmeter (PHTEK, Labitec, Londrina-PR, Brazil), magnetic stirrer (Warmnest, 78HW-1 model, Brazil), UV/Vis spectrophotometer (Even, IL-562 model, Brazil). For data processing, a graphical/statistical program Origin version 6.1 was used (OriginLab Corporation, Northampton, MA 01060, United States).

The Scanning Electron Microscope (SEM) analysis of the vermicompost resulted in a morphological aspect as shown in Figure 2. According to this figure, vermicompost particle surface had heterogeneous and irregular morphology. These morphological characteristics contribute to larger surface area and, consequently, to the adsorption processes [19]. Similar results using SEM for vermicompost were found by [20]. Pereira et al. [10] also verified that the vermicompost, which was efficiently used to adsorb crystal violet and methylene blue dyes, presented very irregular surface.

Figure 3 indicates the presence of chemical groups able to adsorb pollutants chemical species as those evaluated in this study [21]. As can be seen, infrared spectrum of vermicompost revealed characteristic bands, as those associated to N-H stretching (asymmetrical and symmetrical) at 3694 cm⁻¹ and 3619 cm⁻¹ [20] [22]. Vermicompost spectrum was also characterized by a strong and broad band from 3500 - 3000 cm⁻¹ (3522 cm⁻¹ and 3436 cm⁻¹) associated to the stretching u (O-H) of adsorbed water molecules or H-bonded OH groups, including those presents in acid carboxylic, phenolic, and alcoholic groups [20] [23]. The weak absorption at 2900 cm⁻¹ was related to aliphatic C-H stretching, indicating the absence of large amounts of CH₂ and CH₃ groups [20] [23]. Absorption between 1750 - 1500 cm⁻¹ (1640 cm⁻¹) was associated to C-O and C=O stretching bands attributed to the COOH groups, or to C-H of alkene group (Silverstein et al., 1991; Stevenson & Goh, 1971). Adsorption band at 1400 cm⁻¹ was due to OH deformation and C-O stretching of phenolic OH groups, and to C-H deformation of CH₂ and CH₃ groups [22] [23]. The broad and weak band at 1200 cm ⁻¹ is assigned to C-O stretching and OH deformation of COOH. The strong bands at 1002 cm⁻¹ and 909 cm⁻¹ were associated to the C-H stretching of alkene group [22].

Figure 4 reveal that through these characteristics showed in Figure 3 is possible to suppose some interactions between some vermicompost chemical groups and both dyes. The hydrogen bond appears as one of these possible chemical interactions. However, the hydrogen bonds are pH medium dependent. In this experiment we utilized aqueous solutions dyes with pH around 7.0.

As can be seen in Figure 5, mechanical agitation times of 10 and 30 minutes were associated with the higher adsorption percentages of Indigo Blue. It is

important to note that there was no significant difference between adsorptions achieved at 10 and 30 minutes. Thus, the mechanical agitation time of 10 minutes was adopted in all adsorption tests involving Indigo Blue. In turn, as indicated in Figure 5, there was a clear favoring of Congo Red adsorption at 25 minutes, thus justifying the use of this mechanical agitation time in all adsorption tests with Congo Red. Moreover, it was possible to notice desorption events along the time scale indicated in Figure 5. Raymundo et al. [6] also observed similar results, which were achieved when sugar cane bagasse was used to adsorb Congo Red. Ribeiro et al. [24] also noticed that adsorption of Indigo Blue by Ziziphusjoazeiro peel showed desorption events as mechanical agitation time increased.

The influence of different vermicompost masses (g) on the adsorption of Congo Red (40 mg∙L⁻¹) and Indigo Blue (1000 mg∙L⁻¹) is plotted in Figure 6. It was shown that Congo Red adsorption percentage presented small increase from 0.5 to 1.0 g of vermicompost (approximately 96% to 100%). From 1.0 g , adsorption percentage of Congo Red remained practically constant. Therefore, 1.0 g of vermicompost was selected for determining the maximum adsorption capacity for

Congo Red. Indigo Blue adsorption percentage increased when vermicompost mass varied from 0.5 and 3.0 g (approximately 87.7% to 99.9%), remaining unchanged with the addition of larger masses (Figure 6), thus justifying the use of 3.0 g of vermicompost along the tests to estimate the maximum adsorption capacity for Indigo Blue. Raymundo et al. [6] utilized sugar-cane bagasse as adsorbent for treating textile wastewater contaminated with Congo Red (100 μM) to remove approximately 88% of this dye using 10 g of adsorbent. This outcome was considerably different in relation to the vermicompost evaluated in this work (Figure 6). In a work developed by [25], it was determined that 1.5 g of wood sawdust powder from Corymbiacitriodora was able to remove approximately 90% of Congo Red (25 mg∙L⁻¹) from aqueous medium. Ribeiro^A et al. [4] ascertained the liquorice Glycyrrhizaglabra L. root powder as a natural adsorbent to remove textile dyes in water. They demonstrated that 2.0 g of this adsorbent were able to remove 90% and 89% of Congo Red (25 mg∙L⁻¹) and Indigo Blue (1000 mg∙L⁻¹), respectively. Ribeiro et al. [24] demonstrated that 1.5 g of Zizyphusjoazeiro Mart peel removed approximately 84% of Indigo Blue (1000 mg∙L⁻¹) from aqueous medium.

The no linearized adsorption isotherms for both dyes (Figure 7 and Figure 8) were built the ration q (mg∙g⁻¹) against C_eq (mg∙L⁻¹), where q (mg∙g⁻¹) is the adsorbed amount of dye per gram of vermicompost in each dye concentration evaluated, while C_eq (mg∙L⁻¹) is the equilibrium concentration of dyes in the supernatants (dye amount no adsorbed by vermicompost). The forms of this isotherms indicate relatively favorable conditions for adsorption [26]. Mathematically, an adsorption isotherm can be treated by different models, but we adopted

the Langmuir Model [27], which is one the most used models to evaluate the adsorption of pollutants from aqueous media. Langmuir Model considers that adsorption occurs in monolayers, and all adsorption sites have the same energy. Equation (3) is the general equation in which the Langmuir model is based on:

q = K L ⋅ M A C ⋅ C e q ⋅ ( 1 + K L ⋅ C e q ) − 1 (3)

where: q and C_eq have the definitions already given, K_L is the Langmuir constant (L∙g⁻¹), which is related to the adsorption energy, and MAC means maximum adsorption capacity of vermicompost in relation to Congo Red and Indigo Blue, as previously defined. The MAC unit is mg∙g⁻¹ (mg of adsorbed dye per gram of vermicompost).

As can be noted, both adsorption isotherms were not linear (Figure 7 and Figure 8), thus presenting a similarity in relation to the results derived from other studies regarding the use of natural adsorbents and synthetic dyes [4] [6] [10] [24] [25].

Maximum adsorption capacity (MAC) and Langmuir constant (K_L) can be estimated after linearizing Equation (3) [plotting the ration C_eq/q (g∙L⁻¹) against C_eq (mg∙L⁻¹)], thus resulting in linearized isotherms (Figure 9 and Figure 10) described by Equation (4). In this sense, after analyzing Equation (4), MAC could be defined as the inverse of the angular coefficient. In turn, K_L could be estimated as 1 divided by product between MAC and linear coefficient.

C e q / q = ( K L ⋅ M A C ) − 1 + ( M A C ) − 1 ⋅ C e q (4)

When Equation (4) was specifically applied to the data that generated adsorption isotherms showed in Figure 7 and Figure 8, it was possible to obtain Equations (5) and (6), respectively. Equations (5) and (6) represent the lines indicated in Figure 9 and Figure 10, respectively.

C e q / q = 0.229 + 0.0430 ⋅ C e q (5)

C e q / q = 0.1610 + 0.0248 ⋅ C e q (6)

From the K_L values, it was possible to obtain the ΔG_ads concerning the adsorption processes through Equation (7):

Δ G a d s = − R T ln K L (7)

where ΔG_ads is the variations of Gibbs free energy related to the adsorption processes, R is the universal gas constant (8.314472 J∙mol⁻¹∙K⁻¹), T is the temperature in Kelvin (298.15 K), and K_L is a constant related to the adsorption or binding energy (L∙mol⁻¹), between vermicompost and dyes. The values of K_L were originally calculated in L∙g⁻¹, but these values were subsequently converted to L∙mol⁻¹. Table 1 lists all thermodynamic parameters concerning the adsorption of Congo Red and Indigo Blue on vermicompost. As can be seen in Table 1, the R² (Determination Coefficient) values for Equations (5) and (6) highlighted good adjustments of the Langmuir Model in relation to the evaluated adsorption results. The values of ΔG_adsindicate the nature of adsorption. Thus, ΔG_ads < −200 kJ∙mol⁻¹ reveals chemical adsorption with low reversibility degree [28]. In this context, all adsorptions that were evaluated in this study could be classified as

physical adsorption, in which it is observed weaker interactions between dyes and vermicompost. Therefore, it can be said that the chemical interactions shown in Figure 4 are not dominant, occurring discreetly in a much lesser degree than before mentioned.

These discrete values of ΔG_ads are consistent with the structures of the evaluated dyes (Congo Red and Indigo Blue), which have bulky molecular structures with high electronic density. Both features are responsible for steric impediments and electrostatic repulsions, thus avoiding effective accommodation of the dyes on vermicompost particles. Based on the K_L values, which are related to the adsorption energy, it was possible to identify similar affinities of both dyes in relation to the adsorption sites. Ultimately, this outcome reveals that both molecular structures, although remarkably different, presented a set of common characteristics capable of decreasing adsorption strength. Despite low adsorption strength, which are related to the relatively small K_L values, reasonable values of MAC were achieved for both dyes, mainly for Indigo Blue. This finding means that despite the low adsorption energy, appreciable amounts of Congo Red and Indigo Blue could be retained per gram of vermicompost. The comparison between the MAC values obtained in this work with the values shown in Table 2 reveals that the adsorption of CR and IB by natural adsorbents is many varied [4] [6] [17] [24] [25] [29] - [35]. For CR, values vary between 0.52 mg∙g⁻¹ [25] and 751 mg∙g⁻¹ [29]. While for IB all values, except one [4], have MAC around 50 mg∙g⁻¹ [17] [24] [30]. Making a comparative critical analysis, the values revealed in this work are satisfactory. In addition, the vermicompost used in this work is totally natural. No chemical modifications were made to it.

The dimensionless constant (R_L) is another thermodynamic parameter to favoring the adsorption process. In this case, 0 < R_L ≤ 1 indicates favorable adsorptions, while R_L > 1 can be attributed to unfavorable adsorptions (Esmaeili and Beni 2015). R_L is calculated by means of Equation (8), where MAC is the maximum adsorption capacity and C₀ is the highest dye concentration (500 and 4000 mg∙L⁻¹ for Congo Red and Indigo Blue, respectively) used for building adsorption isotherm.

R L = 1 / ( 1 + M A C ⋅ C 0 ) (8)

After substituting MAC and C₀ values in Equation (8), it was possible to achieve the following values of R_L: 0.000086 and 0.0000062 for Congo Red and Indigo Blue, respectively. In this sense, adsorptions of Congo Red and Indigo Blue were favorable.