Introduction

The dyes are often used in the textile, leather, cosmetic, plastic, pharmaceuticals, and food process industries (Garg et al. 2004). The discharges from these industries carry a series of intermediate dyes, which generate large volumes of effluents with different degrees of toxicity (Bhatti et al. 2020). In this way, the colored pollutants are becoming common at hydric bodies, and thus a major preoccupation with the ecosystems (Baig et al. 2019; Wekoye et al. 2020). The methylene blue is the most common dye for its category, which is used in the dyeing process of the cotton and silk (Hameed et al. 2007). In addition to the environmental problems, the strident exposition to the MB causes several health injuries (Vadivelan and Kumar 2005).

Several technologies have been employed for the removal of colored contaminants in water, and among them, the adsorption has been found to be a promising option (Asif et al., 2016) due to applicability at different scenarios, easy implementation, the requirement of less space, and satisfactory cost-benefit when they aligned to efficient adsorbents (Bonilla-Petriciolet et al. 2017; Salleh et al. 2011; Shakoor and Nasar 2018; Rafatullah et al. 2010; Li et al. 2018).

In the literature, it is possible to find the application and development of different adsorbents for the removal of several pollutants. For instance, chitin biochar has been employed for the removal of phenol and 2-nitrophenol (Li et al. 2019a). Annona crassiflora seeds were employed as adsorbents for the removal of crystal violet dye (Franco et al. 2020), while the Azadirachta indica was used to remove Cu2+ and Pb+2 ions in water (Costa et al. 2020). Chitosan/polyamide nanofibers have been used for the removal of Reactive Black 5 and Ponceau 4R (Li et al. 2019b). Brazilian berry seeds (Eugenia uniflora) (Georgin et al. 2020a) and psyllium seeds (Malakootian and Heidari 2018) were applied for adsorption of methylene blue and reactive orange 16, respectively. Besides, the application of carbon nanotubes in single and binary mixtures for the adsorption of rhodamine B and crystal violet (Li et al. 2020a) was studied. Morgina stenopetala seeds were used as adsorbents of Cd2+, Pb2+, and Cu2+ (Kebede et al. 2018), whereas black cumin (Nigella sativa L.) was used for adsorption of methylene blue (Siddiqui et al. 2018). The application of ashitaba waste and walnut shell–based activated carbon for the adsorption of Congo red and methylene blue was also verified (Li et al. 2020b). Last, papaya seeds were applied as potential adsorbents for the removal of different types of pollutants in water (Weber et al. 2013; Paz et al. 2013; Foletto et al. 2013; Weber et al. 2014). Although there are many studies concerning the employment of different seed-based adsorbents, no work using seed residues of Anadenanthera macrocarpa and Cedrela fissilis has been reported so far.

The Anadenanthera macrocarpa is a tree species native to Brazil, belonging to the Fabaceae family, popularly known as red angico. Rich in tannins, its bark is indicated in folk medicine for the treatment of respiratory diseases, inflammatory processes, and healing (Figueredo et al. 2013). Another forest species native to Brazil and of great occurrence in South America is Cedrela fissilis, popularly known as pink cedar, belonging to the Meliaceae family. Its wood is considered noble, with high commercial value, due to this species has been suffering in the last years with illegal deforestation for later commercialization (Muellner et al. 2009; Muellner et al. 2010). Both forest species have peculiarly shaped fruits, with a woody aspect and brown color. Besides, they open at certain times of the year releasing the seeds present in their internal structure, and then fall to the soil, without having other uses, becoming waste (Pennington et al. 2004; Siqueira-Silva et al. 2016).

This work aimed to investigate the potential application of seed residues from Anadenanthera macrocarpa and Cedrela fissilis as adsorbents for the removal of methylene blue in the water. The physical and chemical properties of materials were determined by different characterization techniques. The adsorption assays were conducted at optimum experimental conditions, and kinetic and equilibrium experiments were developed. Besides, the adsorbents were also applied for color removal from simulated textile effluent.

Materials and methods

Chemicals and solution preparation

Methylene blue (MB) (molecular formula: C16H18C1N3S, molecular weight: 319.87 g mol−1, dimensions: 14.3 Å wide, 6.1 Å depth, 4 Å thick, and molecular volume of 241.9 cm3 mol−1) was obtained from Sigma-Aldrich. The MB solutions were prepared using deionized water, according to the desired experimental condition. The pH solution was measured using a Digimed apparatus (DM 20, Brazil). All the adsorption experiments were conducted by a thermostatic agitator (Marconi, MA 093, Brazil) at 150 rpm.

Preparation of adsorbents

The seed residues of Anadenanthera macrocarpa and Cedrela fissilis were collected in the forest at the Federal University of Santa Maria (Brazil). The seeds were dried at 50 °C for 24 h and then comminuted employing a knife grinder (Marconi, MA 340, Brazil) and further dried at 50 °C for 3 h. After, the particles were sieved by a 60-mesh sieve, obtaining particles with a diameter inferior to 250 μm. Both adsorbents presented a brown color, as shown in Fig. S1. The powders were stored in vacuum bags and labeled as RASP (red angico seed powder) and CSP (cedar seed powder).

Adsorbent characterization

It is essential to investigate the adsorbent properties, such as the morphology, functional groups, and crystallographic structure. For the morphological analysis, a scanning electron microscopy (SEM, Tescan - Vega 3SB, Czech Republic) was employed, operating at 10 kV. The identification of the major functional groups for RASP and CSP samples before and after the adsorption was carried out by a Fourier transform infrared spectrophotometer (FT-IR, Shimadzu, Prestige-21, Japan). In this analysis, 35 mg of each adsorbent was pressed with dried analytical potassium bromate. X-ray diffraction (XRD) was used to investigate the structure of the samples (Rigaku, Miniflex 300, Japan).

Batch adsorption experiments

The adsorption assays were conducted to investigate the dosage effect, kinetic, and equilibrium profile. The dosage effect was analyzed using 0.5 to 1.5 g L−1. For this, the adsorbent (RASP or CSP) was put in contact with 50 mL of MB solution (initial concentration of 100 mg L−1). The mixtures were agitated during 180 min at 298 K. The kinetic experiments were realized using the optimum dosage, at the initial concentrations of 25, 50, 100, and 200 mg L−1. Similarly, the adsorbent was added to 50 mL of MB solution at 298 K. The samples were collected at 0, 5, 15, 30, 45, 60, 120, and 180 min. Last, the equilibrium experiments were obtained at 298, 308, 318, and 328 K, with initial concentrations of 50, 100, 150, 200, and 300 mg L−1. The adsorbents were mixed with 50 mL of MB solution and agitated for 5 h to ensure the equilibrium.

The detection of the MB concentration in the liquid phase was done by employing a UV-vis spectrophotometer (Shimadzu, UV mini 1240, Japan), operating at 664 nm. All tests were realized in triplicate (n = 3), and the blank test was also carried out to guarantee the data reproducibility. All samples were separated by centrifugation (Centribio, 80-2B, Brazil) at 4000 rpm for 20 min. The percentage of dye removal, adsorption capacity at any time, and equilibrium were used to evaluate the adsorption process.

Kinetic, isotherm, and thermodynamic models

The adsorption kinetics of MB dye on the adsorbents was analyzed according to the Elovich (Elovich and Larinov 1962), pseudo-second-order (Ho and Mckay 1999), pseudo-first-order (Lagergren 1898), and general order models (Liu and Shen 2008) (equations shown in Supplementary material (S.2)). For the adsorption isotherms, Langmuir (Langmuir 1918), Freundlich (Freundlich 1906), and Tóth (Tóth 2002) models were chosen (Equations are shown in Supplementary material (S.3)). The thermodynamic calculations were based on the methodology proposed by Lima et al. (2019), whose equations are shown in Supplementary material (S.4). The fit quality evaluation is shown in Supplementary material (S.5). The model parameters were estimated through script programming on Matlab 2017.

Preparation of synthetic textile effluent

Some studies report the preparation of synthetic wastewater containing textile dyes, from the mixture of dye with a source of water, and chemicals. The mixture aims to combine the residual features with the particularities of the actual effluents, which contains several chemicals, auxiliaries, and dyes added during the textile production process (Yaseen and Scholz 2018). Thus, a simulated effluent was prepared to verify the capacity of angico and cedar samples for the treatment of residual effluents containing organic pollutants and salinity (Table 1). For each test, 4.5 g L−1 of adsorbent was added in 100 mL of simulated effluent solution and stirred at 200 rpm for 4 h under room temperature (298 K). The dosage of 4.5 g L−1 was applied in this work due to the unsatisfactory results obtained by the using dosage equal to 1 g L−1, as observed in preliminary tests realized by our research group (Georgin et al. 2020c). Then, absorbance scans of the solution were performed before and after adsorption tests, at a wavelength of 200 to 800 nm, using the UV-Vis spectrophotometer (Shimadzu, UV-2600, Japan). The color removal was determined by the ratio between the area of the curves using the Matlab 2017 with trapz function.

Table 1 Chemical composition and concentration of the simulated textile effluent
Full size table

Results and discussion

Characteristics of RASP and CSP samples

The SEM images with different magnifications of RASP and CSP samples are shown in Fig. 1. The surface of the RASP sample is highly rugous, irregular, and heterogeneous, with the presence of empty spaces smaller than 10 μm (Fig. 1a, b). Likewise, the CSP sample exhibits irregularities on its morphology, with a rough surface containing cavities. It is of interest that the materials presented such characteristics, since these empty spaces and channels may positively favor the adsorption (Georgin et al. 2018a, b). Similar characteristics were found in other lignocellulosic materials such as Pará chestnut husk (Georgin et al. 2018a, b), Araticum (Annona crassiflora) seeds (Franco et al. 2020), and golden trumpet tree bark (Hernandes et al. 2019).

Fig. 1
figure 1

SEM images for the RASP (a, b) and CASP (c, d), with magnifications of × 200 and × 5000

Full size image

Figure 2 shows the FT-IR spectra of RASP and CSP samples. The forest wood (seeds, wood, leaves) are lignocellulosic materials mainly constituted of three groups: lignin, cellulose, and hemicellulose (Salleh et al. 2011). In the case of the RASP and CSP samples, the following bands were identified: the high stretch at 3434 cm−1 is associated with the presence of OH (Saha et al. 2012; Feng et al. 2011); the band deformation at 2917 cm−1 is attributed to the C–H bond (alkyl, aliphatic and aromatic groups); at 1731 cm−1, on the CPS sample, can be found a C=O bond (Babalola et al. 2016); band at 1664 cm−1 can be related to C=O and N–H bonds, belonging to the amide (Kumar and Ahmad 2011); the RASP sample presents a 1535 cm−1 minor band related to the C=O (ketone or carbonyl) (Chakraborty et al. 2011). Last, the bands at 1050 and 605 cm−1 are related to C–OH, C–O–C, and aromatic bonds (Román et al. 2013). From Fig. 2, it was found that after the adsorption, the transmittance increased for both adsorbents. In this case, the formation/absence of bands was not found in comparison with the original adsorbent samples. The general increase of the transmittance suggests that the main functional groups of the material are involved in the adsorption mechanism. Groups such as O–H and CO may be involved in the adsorption; this is related to their negative nature, leading to interact with positively charged MB molecules (Guo et al. 2020; Pang et al. 2019).

Fig. 2
figure 2

FT-IR spectra and respective identified bands of RASP and CSP samples

Full size image

The XRD patterns for the RASP and CSP samples are shown in Fig. 3. Both materials exhibit a broad peak, with a maximum at 26.39 and 22.19° for the RASP and CSP samples, respectively. The wide peaks from 10° to 30° can be attributed to the lignin content (Georgin et al. 2020b). It is known that an adsorbent with amorphous characteristics has a more disorganized structure and larger void spaces, and these characteristics contribute to the adsorption of dye molecules (Georgin et al. 2020b). Similar behaviors were observed by the XRD analysis of the bark of Moringa oleifera (Reddy et al. 2011).

Fig. 3
figure 3

XRD patterns of RASP and CSP samples

Full size image

pH justification and adsorbent dosage effect

The pH plays an important role in the adsorption process of the dye molecules since it controls the electrostatic force intensity of charges transmitted by dye molecules (Mahmoud et al. 2012). In this study, all tests were performed at the natural pH of the MB solution (pH = 8). Several studies prove that the pH around 8 is ideal for MB removal in water (Franciski et al. 2018; Netto et al. 2019), and the explanation is that at low pH, there is a strong repulsion between the positively charged dye ions and the negatively charged adsorbent surface, due to protonation of the functional groups present on the surface, resulting in low efficiency in MB removal. The reverse process occurs with an increase in pH value, causing a greater increase in removal efficiency due to deprotonation of the groups that are positively charged on the adsorbent surface, plus the electrostatic forces that attract the negatively charged places in the biosorbent and the dye cations (Saha et al. 2012).

The adsorbent dosage (D) used to remove the dye is also important for determining the adsorption capacity (q) given the initial concentration of the dye in solution (Singh et al. 2017). For both materials, it was found a typical tendency; i.e., the increase of the adsorbent dosage decreased the adsorption capacity and increased the removal percentage, as shown in Fig. 4. In the case of the RAPS sample, the adsorption capacity decreases from 105 to 60 mg g−1 with a removal percentage ranging from 55 to 100%. The values for the CPS sample trend 115 to 54 mg g−1 with the removal of 55 to 100%. It is possible to consider that 1 g L−1 is the optimum condition for the adsorption (intersection of the curves). A similar trend was observed by Jain and Gogate (2018) by using Prunus Dulci bark activated to remove the Acid Green 25 dye.

Fig. 4
figure 4

Dosage effect for the adsorption of MB onto a RASP and b CSP samples

Full size image

Adsorption kinetics and fitted models

The kinetic data for both systems with different initial concentrations (25, 50, 100, and 200 mg L−1) are shown in Fig. 5. The first aspect to be noticed is that the initial concentration influenced the inclination of the adsorption data. About the equilibrium time for both systems, it is reached at approximately 60 min. However, for the cases where the C0 is 25 mg L−1, the equilibrium time is reached at 30 min. For the adsorption capacities, the RASP sample reached 174.57 mg g−1, and the CSP sample reached 162.88 mg g−1 in the higher initial concentration (200 mg L−1). The behavior of adsorption kinetics is due to the higher adsorption rate that occurs in the first minutes. At this moment, the surface of the material possesses more adsorption empty sites. As time increases, the adsorption rate decreases as well since the adsorption site is more occupied and reaching the equilibrium. A similar kinetic profile was also observed by Aichour and Zaghouane-Boudiaf (2019), using composites of modified citrus peel based on cellulose and calcium alginate for the removal of cationic dyes Gentian violet and methylene blue.

Fig. 5
figure 5

Kinetic curves for the adsorption of MB onto a RASP and b CSP samples

Full size image

The estimated parameters for the kinetic models related to RASP and CSP samples are shown in Tables 2 and 3, respectively. From all the evaluated models, the pseudo-second-order model was the most suitable for describing both systems. For the RASP sample, the pseudo-second-order model resulted in an R2adj ≥ 0.9516 and MSE ≤ 68.17 (mg g−1)2, while the CSP sample achieved R2adj ≥ 0.9732 and MSE ≤ 23.15 (mg g−1)2. Moreover, the pseudo-second-order model was also able to predict the adsorption capacity for all cases and both systems. Kinetic adjustments for the pseudo-second-order model have also been found in the adsorption of orange solimax dye onto malt bagasse (Fontana et al. 2016), adsorption of tartrazine dye onto wood industry residues (Banerjee and Chattopadhyaya 2013), and on the adsorption of Orange G dye by modified Pyracantha coccínea (Gorgulu and Celik 2013).

Table 2 Kinetic parameters estimated for the adsorption of MB onto RASP adsorbent
Full size table
Table 3 Kinetic parameters estimated for the adsorption of MB onto CSP adsorbent
Full size table

Adsorption isotherm and thermodynamic estimation

The adsorption equilibrium for the MB adsorption onto RASP and CSP samples is shown in Fig. 6. At first observation, both systems show increasing value for the adsorption capacities with the rise of the temperature. For the RASP sample, the equilibrium adsorption capacity change from 205.92 to 227.57 mg g−1. The CSP sample shows a similar trend with a slight increase of 202.78 to 230.06 mg g−1. This indicates an endothermic nature for both the systems. In the adsorption process, the temperature has two important effects. The increase in temperature decreases the viscosity of the solution, which increases the diffusion rate of adsorbate molecules through the outer boundary layer and within the pores of adsorbent particles, thus increasing the adsorption sites, and consequently increasing the adsorption capacity (Al-Qodah 2000; Nandi et al. 2009; Wekoye et al. 2020).

Fig. 6
figure 6

Equilibrium curves for the biosorption of MB onto a RASP and b CSP samples

Full size image

The computed parameters for the isotherm models are presented in Tables 4 and 5. For both cases, the Tóth model shows better conformity for describing the experimental data. The goodness of fit shows R2adj ≥ 0.965 and MSE ≤ 145.59 (mg g−1)2 for the RASP and R2adj ≥ 0.990 and MSE ≤ 57.02 (mg g−1)2 for CSP. The Tóth model is an expansion of the Langmuir model, that considerate multi-layer adsorption and the heterogeneous surface of the materials (Wu et al. 2013). Thus, the model suggests that the MB adsorption process onto the RASP and CSP adsorbents may be heterogeneous. The Tóth model was able to describe other liquid systems, such as adsorption of nickel(II) onto biochar derived from the walnut shell (Georgieva et al. 2020). Another key thing to remember, this model has the greatest affinity to surfaces with non-uniform distribution of the active sites, leading to random adsorption in several layers (Chandra and Chattopadhyay 2019).

Table 4 Equilibrium parameters estimated for the adsorption of MB onto RASP adsorbent
Full size table
Table 5 Equilibrium parameters estimated for the adsorption of MB onto CSP adsorbent
Full size table

A comparison among adsorption capacities of RASP and CSP samples with other adsorbents reported in the literature for MB removal is presented in Table 6. The results reveal that the samples prepared in this work have remarkable adsorptive performances for MB in aqueous solution. Therefore, these residues become alternative adsorbents for the removal of dyes from aqueous solutions, assuming the following advantages: abundance and easy availability in nature, and thus being ecological.

Table 6 Comparison of RASP and CSP adsorbents with other adsorbents employed for the removal of MB
Full size table

The thermodynamic computed values are presented in Table 7. The equilibrium constant (Ke) was estimated from the Tóth isotherm. It was found that the adsorption process is spontaneous, with ΔG0 ranging from − 30.07 to − 34.10 kJ mol−1 for the RASP sample and − 31.61 to − 34.93 kJ mol−1 for the CSP sample. The ΔH0 for both systems was found 9.87 kJ mol−1 (RASP) and 1.30 kJ mol−1 (CSP), indicating an endothermic behavior, which corroborates the isotherm profile for both systems. Last, the ΔS0 is related to the adjustments of the molecules on the material surface and, in this work, the values were 0.133 and 0.11 kJ mol−1 K−1 for the RASP and CSP samples, respectively.

Table 7 Thermodynamics parameters for the RASP/MB and CSP/MB systems
Full size table

Color removal from simulated textile effluent

In general, the adsorption studies in the batch system are focused on the removal of a single compound. However, this work investigated the application of the RASP and CSP samples for the color removal of simulated effluent (Table 1). The UV-Vis absorption spectra for the effluents before and after the application of the RASP and CSP samples are shown in Fig. 7. From the areas under the curves, it is possible to estimate that the RASP sample was able to remove 74.58% of color, and the CSP sample, 78.54%. According to the results observed in the UV region, it is possible to infer that there was the formation of other compounds due to the breaking of the covalent bonds present on the dye molecule, which are simultaneously removed by the adsorbents.

Fig. 7
figure 7

UV-Vis absorption spectra before and after the effluent treatment with RASP and CSP samples

Full size image

Conclusion

The residue-based adsorbents investigated in this work (Anadenanthera macrocarpa and Cedrela fissilis) were effective in the removal of methylene blue dye in water. From the characterization techniques, it was possible to find chemical groups typical of lignocellulosic materials. Both materials had a rough, highly heterogeneous surface with the presence of particles with varying shapes and sizes. The dosage of 1 g L−1 of RAPS and CSP was enough to remove 80% and 83% of the MB dye. The adsorption kinetics for both materials achieved the equilibrium time at 60 min when the concentration was higher than 50 mg L−1. The pseudo-second-order model was the most suitable for describing the adsorption kinetics. From the equilibrium experiments, it was found that the Anadenanthera macrocarpa presented a maximum equilibrium capacity of 228 mg g−1 and Cedrela fissilis, a similar capacity of 230 mg g−1 at 328 K. The Tóth model is the most proper one for describing the isotherm data for both systems and indicates that the adsorption can occur in multi-layer and/or the surface of the RASP and CASP is heterogeneous. In the treatment of the simulated effluent containing different dyes, the materials showed an excellent decolorization performance, removing 74% for the RASP sample and 78% for the CSP sample. Therefore, the seed wastes of angico and cedar become alternative adsorbents for the treatment of textile effluents, presenting as advantages the low cost and the efficiency in removing dyes from aqueous solutions.