Removal of Methylene Blue from Aqueous Solutions by Using Nance (Byrsonima crassifolia) Seeds and Peels as Natural Biosorbents

Instituto Politécnico Nacional, Escuela Nacional de Ciencias Biológicas, Departamento de Biof́ısica, Carpio y Plan de Ayala S/N Santo Tomás 11340, Ciudad de México, Mexico Instituto Politécnico Nacional, Escuela Nacional de Ciencias Biológicas, Departamento de Ingenieŕıa Bioquı́mica, Carpio y Plan de Ayala S/N Santo Tomás 11340, Ciudad de México, Mexico Facultad de Ingenieŕıa de Alimentos, Fundación Universitaria Agraria de Colombia–Uniagraria, Calle 170 # 54a–10, C.P. 111166, Bogotá, Colombia Instituto Politécnico Nacional, Escuela Nacional de Ciencias Biológicas, Departamento de Investigación, Carpio y Plan de Ayala S/N 11340, Ciudad de México, Mexico


Introduction
Contamination of effluents greatly impacts ecosystems frequently causing eutrophication and adverse effects on human health including digestive, skin, cognitive, and vision. Methylthioninium chloride or methylene blue (C 16 H 18 ClN 3 S) (MB) is a synthetic dye that has a thiazine structure, is a cationic compound also known as Swiss blue, and is a contaminant frequently found in urban and industrial discharges due to commercial activities [1]. MB is used in pharmaceutical, personal care, printing, leather, and textile industries and is frequently discharged to sewage from domestic households as well as from small-and medium-sized businesses. It is also employed in medicine to treat methemoglobinemia [2] and exhibits antioxidant, antimalarial, antidepressant, and cardioprotective properties. On the contrary, MB has various harmful effects on human health such as headache, shortness of breath, and high blood pressure [3,4].
Bioremediation of effluents has received a great deal of attention, and the removal of harmful contaminants as metal ions has been extensively studied by using bio-based adsorbents (mainly agricultural residues) as an alternative to the use of other chemical adsorbents. For instance, waste biomass of Phaseolus vulgaris has been used for the removal of lead (II) ions [5], and its husks were used as adsorbent for Cr (VI) removal [6,7]. Agave biomass has been employed as a biosorbent for lead (Pb II) and copper (Cu II) from contaminated water [8]. Additionally, soybean and rice hulls were utilized for the adsorption of Hg from dilute aqueous solutions [9]. Also, the adsorption of Cd (II) was carried out by using citrus peels in a packed bed column [10], and the removal of basic fuchsin from water was carried out with mussel powdered eggshell membranes as the adsorbent [11]. e removal of MB by using biosorbents has also received attention. e adsorption of MB on Citrus limetta peel waste was studied with special focus on the associated costs and on the experimental fitting of data to various isotherm models [12]. Yan et al. [13] reported the sorption of methylene blue by using carboxymethyl cellulose, finding that the adsorption was dependent on pH, but independent of the temperature, and found that the Langmuir isotherm best fitted the experimental data. Aryee et al. [14] proposed the functionalization of peanut husk with iminodiacetic acid for the removal of MB from different solutions and reported mass transfer implications of the process. e adsorption of methylene blue on carbonaceous hydrochar derived from coffee husk waste as well as corresponding kinetics was also reported by Tran et al. [15]. Alver et al. [16] reported the adsorption of MB on alginate/rice husk biocomposite beads and found that pH did not influence sorption efficiency.
Nance fruit is consumed in tropical and subtropical countries and regions as in México and Central and South America, is a source of functional compounds [17,18], and has been used at commercial levels to prepare soft drinks, ice cream, salads, and liquors [19]. Seeds and husks represent approximately 3 to 14% of the fresh fruit and are usually composted or disposed, and no other use has been given to these materials [19]. e seeds and peels of fruits of the family Malpighiaceae are rich in dietary fiber [20] which makes them potentially attractive biosorbents for cationic compounds such as methylene blue provided the charge balance on the adsorbents is negative [20]. Actually, there is one published report in which nance seeds were used as biosorbents for Cd (II) and Pb (II) in aqueous solutions, achieving percentages of removal up to 84% for Cd (II) to pH 8 and of 82% for Pb (II) to pH 5 [21]. Given the chemical characteristics of MB and nance seeds and peels, these materials are potential biosorbents for the removal of MB from aqueous solutions at different pH values. Also, the usage of these materials would aid in the revalorization of food processing byproducts. e aim of this work was to study the possible use of seeds and peels of nance as biosorbents for the removal of methylene blue from aqueous solutions and to infer on chemical interactions between the dye and the biosorbents by means of isotherms' modeling and spectroscopic and image analyses of the materials before and after MB adsorption.

Raw Materials.
Fresh nance was acquired from the local market of Iguala, Guerrero, México. Fruits suitable for consumption were chosen according to the Mexican standard NOM-120-SSA1-1994. Even yellow color of the skin and a size ranging 1.5-2.0 cm in diameter were set as the criteria of commercial maturity [22]. Selected fruits were washed in running water, immersed in a 100 ppm sodium hypochlorite solution for 3 minutes, and rinsed with running water and finally with distilled water as reported by Díaz [22].

Separation of the Pulp and Collection of Husks and Seeds.
Nance fruits were fed to a pulper machine (POLI, DERE-1, 022.301107.09, México) fitted with a 0.1 mm mesh to separate the pulp from seeds and peels which were collected and oven-dried ( elco, Precision Scientific Company, model 19) until a 4 ± 1% moisture content (mc) wet basis was reached [23][24][25]. Dried samples were grinded by using a laboratory mortar and sieved to 60-200 mesh particles.

Point of Zero Charge (pzc).
e point of zero charge was determined according to Albis et al. [20]. In brief, 50 mL of 0.01 M NaCl was adjusted to pH from 2 to 12 at 1 pH unit interval by using 0.01 M NaOH and HCl. en, 0.1 g of the sorbents was added, and the mixture was stirred for 48 h. pH of each batch was measured (pH meter: Hach Sension 1, model 51700-23, China). Initial and final pH values were recorded and plotted. e pzc was found as pH of the intersection of plotted curves with a 45°straight line in the same graph [26].

Methylene Blue Adsorption Experiments.
Adsorption experiments were carried out by preparing 10 MB solutions at different concentrations (25,50,75,100,125,150,175,200,225, and 250 ppm) [3,8,27]. en, 0.1 g of the adsorbents (nance seeds and peels) was added, and pH was adjusted to 7 and 10 to simulate urban and industrial (textile) effluents [28]. Next, the systems were stirred for 4 h and filtered by using Whatman No. 40 filter paper. e filtrate was collected, and absorbance at 664 nm was measured and interpolated in a standard curve to finally determine the adsorbed mass of MB as follows: in which q is the amount of adsorbed MB per unit mass of the adsorbent (mg/g), C 0 is the initial concentration of MB (mg/L), C e is the final concentration of MB (mg/L), V is the volume of MB solution adsorbed (L), and m is the mass of the sorbent (g). e percentage of MB removal or efficiency of adsorption (η) was obtained as follows: Modelling of methylene blue adsorption is shown as follows.
Experimental adsorption data were fitted to Freundlich and Langmuir isotherms [29].
e Freundlich model is in which q is the amount of adsorbed MB per unit mass of the adsorbent (mg/g), C e is the concentration of MB at equilibrium (mg/g), K F is Freundlich constant ((mg/g) (L/mg) 1/n ), and n is Freundlich intensity constant (dimensionless). In its logarithmic linear form, Freundlich isotherm is By linear regression, the values of K F and n can be obtained from the intercept and slope, respectively. e Langmuir isotherm was applied to the sorption data as follows: In its linear form, in which q is the amount of adsorbed MB per unit mass of the adsorbent (mg/g), K L is Langmuir constant (L/mg), Q max is the maximum adsorption (mg/g), and Ce is the residual concentration of the adsorbate (mg/L). e Langmuir isotherm may be represented by the dimensionless equilibrium parameter (R La ) as given by Saygılı et al. [29]: in which K L is Langmuir constant (L/mg) and C 0 is the initial concentration of methylene blue (mg/L). Values for R La indicate various nature of isotherms such as irreversible (R La � 0), favorable (0 < R La < 1), linear (R La �1), or unfavorable (R La > 1). Efficiency of adsorption η was evaluated in all cases as [8] η in which C i � and C f are the initial and final concentrations of MB in the adsorbents.

Mass Transfer in Biosorption.
e individual mass transfer coefficient for the liquid phase (k) at the onset of adsorption may be described by the rate equation [30]: (9) in which N is the mass transfer flux (mg/h), k i is the initial mass transfer coefficient (s −1 ), x b is the concentration of methylene blue in the bulk phase (liquid) at the beginning of the experiment, and x 0 is the concentration of methylene blue in the adsorbent (mg/g). N can be obtained by dividing the amount of methylene blue adsorbed for each concentration of the colorant in the liquid by the four hours of duration of the adsorption, x b can be considered as the initial concentration of the dye, and x 0 may, in every case, be considered as zero since at the beginning of the adsorption, the surface of the adsorbent is free of colorant.

Biosorption ermodynamics.
To investigate the change in Gibbs free energy, the following equation was applied [29]: in which ΔG 0 is the change in Gibbs free energy (KJ/mol), R is the universal gas constant, T is the absolute temperature (K), and K L is Langmuir constant (dimensionless).

Fourier-Transform Infrared Spectroscopy (FTIR).
Nance seeds and peels were analyzed by FTIR [31] to detect functional groups before and after adsorption of MB. A Frontier FT-IR/NIR PerkinElmer Spectrometer, USA, fitted with a monolithic crystal was used. e spectral range varied from 4000 to 800 cm −1 .

Scanning Electron Microscopy and Image Analysis.
Images of seeds and peels before and after methylene blue adsorption were obtained by scanning electron microscopy (SEM) (Quanta FEG 3D) (500x, 1500x, and 5000x, 15 kV, 26 pixels/nm 2 ) [32]. Fractal dimension (FD) was obtained by using ImageJ v. 1.53g software (National Institutes of Health, USA). Brightness and contrast for all images were adjusted automatically before performing the FracLac analysis. e size and greyscale values were standardized in all images using a calibrated greyscale. e fractal analysis was performed with the FracLac plugin in ImageJ software [33].

Statistical Analyses.
Statistical analyses were carried out by means of an analysis of variance (ANOVA) by applying a significance level α � 0.05. All calculations were carried out by using Minitab 17.1.0 software, and results were expressed as mean values ± standard deviation.

Proximal Analysis.
e proximal analysis of nance seeds and peels is shown in Table 1. It was possible to observe that both seeds and peels had substantial amounts of crude fiber which is their major component (19.85 and 19.94% for peels and seeds, respectively), making these materials attractive as potential biosorbents of a basic dye such as MB [29]. Even though there were no reports found on the composition of nance seeds and peels, it has been reported that the seeds of acerola cherry which belongs to the same family as nance (Malpighia) are a good source of dietary fiber [20].

Point of Zero Charge.
e point of zero charge (pzc) of peels and seeds of nance was 3.42 and 5.96, respectively (Figure 1), and in this regard, Espinoza Rodríguez et al. [21] reported a pzc of 6.0 for nance seeds of an unknown variety. Obtained values of pzc indicated pH at which the materials had a zero net charge of the total particle surface as related to the media [34]. When sorption was carried out above pH values corresponding to the pzc, the negative sites on the surface of seeds and peels of nance will be exposed [35] and will present affinity by methylene blue if media pH is above the reported pKa for MB of 3.8 which favors the predominant cationic forms of this dye [36] that would be attracted to the negatively charged sites on the surface of the adsorbent in a way that efficient adsorption may be carried out at pH above 3.8.
us, adsorption experiments were carried out at pH 7 and 10 for both bioadsorbents simulating urban and industrial (textile) effluents [28]. Figures 2-5, and their corresponding constants are given in Table 2. It was possible to observe that Langmuir and Freundlich isotherm models fitted well (R 2 > 0.99 in all cases) to experimental data. Langmuir isotherm assumes ideal energetically homogeneous active sites and the consequent formation of a monolayer which leads to the assumption of a rather homogeneous structure of the adsorbent [29]. ese considerations could well apply to MB adsorption onto cellulose and lignin which have polymeric structures [37], and the active sites are distributed rather evenly on the adsorbents, leading to rather homogeneous affinity for a comparatively small molecule as that of MB. Saygılı et al. [29] reported that Langmuir isotherm explained the adsorption of MB on dried grape pulp obtained after extraction of the juice. Similar values of K L were obtained by Al-Anber et al. [38] when studying iron (III) adsorption on olive cake, by Alhujaily et al. [39] when removing anionic dyes from aqueous solutions using spent mushroom waste, and by Berber-Villamar et al. [40] who studied the adsorption of azo dye Direct Yellow 27 on corncob from aqueous solutions. It is noteworthy (Figures 2-5) that the slope of Freundlich and Langmuir isotherms corresponding to pH 10 was slightly larger than that at pH 7. is was due to the relatively higher affinity between adsorbent and MB at pH 10, discussed in Section 3.4. Yan et al. [13] reported that the sorption of methylene blue on carboxymethyl cellulose was dependent on media pH. Additionally, all obtained values of the equilibrium parameter (R La ) were lower than 1 ( Table 2) which indicated the favorable nature of the Langmuir isotherm for the sorption process [29].

Sorption Isotherms. Methylene blue sorption isotherms are shown in
Freundlich isotherm is empirical and considers a nonuniform adsorption and a heterogeneous sorption heat distribution during adsorption [29]. e exponent n suggests the nature of the sorbent's surface. Values lower than 1 but larger than 0 indicated a less heterogeneous surface of nance seeds and peels at both pH values (7 and 10), being slightly more favorable at pH 10. is was expected given the higher-affinity adsorbents-MB at this larger pH value as shown in Figures 6 and 7 [41], given its remoteness to pKa as described in Section 3.2. Also, it is suggested that, at values of 1/n ≪ 1, the surface is energetically heterogeneous than for 1/n figures closer to 1. erefore, as 1/n for both materials was, in average, 0.99 and 0.96 for seeds and peels, respectively (Table 2), the adsorption of MB onto seeds and peels of nance took place on rather energetically homogeneous surfaces [41], being slightly more heterogeneous for peels than for seeds. Physical roughness was also found to be slightly more pronounced (higher fractal dimension) for peels than for seeds (Section 3.8). Figures 6 and 7. It was possible to observe that, at pH 7, adsorption reached maximum values of 99 and 98% at MB concentrations of 75 and100 mg/L for peels and seeds, respectively. For the region of low initial concentrations of MB (25-75 mg/L), low adsorption efficiencies were obtained (Figures 6 and  7) [20]. is was due to the small amounts of dye molecules present in the media associated with low activity coefficients in the diluted solutions, causing small amounts of MB deposits on the active sites of the adsorbent. At high initial concentrations of MB, the efficiency remained rather constant, and the small decrements observed for seeds at pH 10 could be due to

Biosorption ermodynamics.
Change in Gibbs free energy was evaluated according to equation (10). Results of ΔG 0 obtained for the two sorbents at the two experimental pH values (7 and 10) are shown in Table 2. All values of ΔG 0 were negative which indicated that the biosorption process was spontaneous within the experimental conditions used, and these values were larger than −80 but lower than 0 kJ/ mol which indicated that the MB adsorption process was due to physical interactions which are a characteristic of physical adsorption [29]. Larger ΔG 0 figures are associated to covalent binding between the adsorbent matrix and the adsorbate and are typical of chemisorption [38].

Bulk Mass Transfer Coefficients.
Average bulk mass transfer coefficients corresponding to the onset of the adsorption are shown in Table 3. No significant differences were found amongst all mass transfer coefficients of both sorbents and pH values which indicated that both adsorbent systems had the same hydrodynamic and driving forces, as well as similar MB-adsorbent affinities at the beginning of the process, given that the surface of nance seeds and peels had irregular surfaces (Sections 3.3 and 3.8) and were free of dye at the onset of the adsorption. Obtained values of k were similar to those reported in [43] for the removal of phenol from wastewater using a forest waste. Knowledge of mass transfer coefficients and isotherms would greatly facilitate the design of an adsorbent system as well as the prediction of the optimum contact time between MB and the biosorbents [30]. Mass transfer to the solid includes transport of the solute from the bulk solution to the film around the adsorbent or film diffusion, internal diffusion from the film to the pores of the adsorbent, and adsorption on the external surface of the adsorbent on specific active sites [43]. Physical adsorption is practically instantaneous so that equilibrium is expected among the surface and the fluid at each location in the adsorbent particles [30]. Figures 8-11, the FTIR spectra of bioadsorbents before and after MB removal are presented, and in Table 4, a summary of main bands detected as well as the type of bond interactions is shown. Both biosorbents presented bands as the O-H stretching at 3290 cm −1 corresponding to polysaccharides, lignin, and phenolic groups [44]; bands at 2924-2923 cm −1 and 2854 cm −1 which are characteristics of the C-H symmetric and nonsymmetric stretching, respectively, present in methyl and methylene groups of lipids, carotenoids, carboxylic groups, and lignin [45]. e decrement in peak intensities between 1370 and 1318 cm −1 corresponded to -OH, -CH 2 , and -CH 3 . Also, the C�O band at 1734 cm −1 gave place to two smaller peaks at 1743 and 1715 cm −1 that corresponded to interactions with  Values ± standard deviation. * Different superscript letters in the same column indicate significant differences (p ≤ 0.05).   Figure 11: FTIR spectrum of nance peels before and after MB adsorption (pH 7). Blue tags correspond to functional groups of MB. Orange tags correspond to functional groups of the peels.  Figure 12: SEM images of the microstructure of adsorbents: peels (a) and seeds (b) before and after MB adsorption (1500x, 5000x, 15 kV). FTIR spectra at pH 7 and 10, and the presence of methylene blue after the adsorption was evident, confirming the suitability of nance seeds and peels to be used as biosorbents for the removal of MB from aqueous solutions.

Microstructure of Bioadsorbents and Image Analysis.
SEM images (1500 and 5000x) of seeds and peels of nance before and after the MB adsorption are shown in Figure 12. Irregular surfaces were found in both biosorbents having fractal dimensions between 1.4 and 1.55 (Figure 13), and peels had a rougher structure (higher FD) than seeds. ese findings agreed with values of constant n of the Freundlich isotherm as discussed in Section 3.3. In the case of peels at 1500x, it was possible to observe that the material without MB was more irregular (higher FD) (p ≤ 0.05) than after the adsorption for which smoother structures could be observed which might be due to MB molecules covering the surface of the adsorbent [29], whereas seeds did not show significant differences (p > 0.05) before and after the adsorption. In the case of images at 5000x, seeds showed more homogeneous structures before than after the adsorption which might be due to scatter adsorption mechanisms during the formation of a monolayer.

Conclusions
Langmuir and Freundlich isotherms appropriately described the sorption of MB, and associated efficiencies were larger than 98% in all cases, being slightly lower for pH 7 due to the closeness with the point of zero charge (pzc) of seeds and peels of nance. At pH 7, adsorption reached maximum values of 99 and 98% at MB concentrations of 75 and 100 mg/L for peels and seeds, respectively. In all cases, Langmuir adsorption was favorable (R La < 1) and spontaneous (negative Gibbs free energy of adsorption). e MB removal process was due to physical interactions, a characteristic of physical adsorption. No significant differences were found amongst bulk mass transfer coefficients for both adsorbents indicating that both adsorbent systems had similar hydrodynamic and driving forces under the conditions of the experiment. Adsorption of MB on the adsorbents was corroborated by FTIR spectroscopy and was mainly due to the presence of active sites distributed across complex polysaccharides in the biosorbents. e structure of adsorbents was evidenced by scanning electron microscopy and image analysis which indicated that both adsorbents had fractal and rougher structures, being more irregular in the peels than the seeds.

Data Availability
All the data used to support the findings of this study are available at Instituto Politécnico Nacional, Escuela Nacional de Ciencias Biológicas, Departamento de Biofísica, México.

Conflicts of Interest
e authors declare no conflicts of interest.