Biosorbent preparation and characterization

The sterile bracts of Araucaria cones were collected from sales points around the town of Curitibanos, located in upland areas of Santa Catarina state in the south of Brazil. The material was milled by a Wiley mill and sieved to granulometry between 20 to 60 mesh. The composition of the sterile bract raw (SR) was characterized in order to determine ash, total extractives and lignin contents (^{Sjöström and Alén 1999}). The moisture was determined in an oven at temperature 103-105 ºC until constant weight; the ash content was made in an oven at 600 ºC for 2.5 h. The total extractive was obtained using a Soxhlet apparatus. Sequential extractions steps were of 6 h each with ethanol:toluene (1:2, v/v), ethanol and dichloromethane, and washed with warm water at the end. Lignin was quantified by the Klason method (^{Nicholson et al. 2014}). Briefly, the extractive-free sample was treated with 72 % sulphuric acid and subsequently heated with dilute acid to hydrolyze the polysaccharides; after that, the solid residue was washed, dried and weighed. These experiments were made in duplicate. In order to remove soluble phenolic compounds present in the SR, it was boiled in distilled water for 2 h, dried in an oven and put in a desiccator until its use; this material was henceforward called SB. The boiling step was needed to prevent soluble compounds from moving into the solution during the bath experiments (^{Michelon et al. 2012}). The most important chemical groups present in the SR and SB were analyzed by Fourier transform-infrared spectroscopy (FTIR) (ABB, FTLA 2000) and thermogravimetric analysis was used to explore the behavior of thermal biosorbent (Shimadzu, TGA-50). The point of zero charge (PZC) was determined by the salt addition method (^{Dahri et al. 2016}) using 0.1 mol/L NaCl. The pH of each salt solution (pHi) was adjusted to a range of 3-10 using 0.1 mol/L HCl and 01 mol/L NaOH. These solutions (50 mL) were mixed with biosorbent (1 g) and agitated until pH stabilization (pHf). The initial pH (pHi) and final pH (pHf) were measured and a diagram was created with the pHi versus pHi-pHf values. The PZC value was obtained from this diagram at the point where the data curve intercepts the abscissa and the value of dpH becomes equal to zero.

Biosorbent characterization

The composition of SR was as follows: ash 1.88 %, total extractives 6.06 %, lignin 45.38 % and holocellulose (hemicellulose + cellulose) 46.68 % (calculated by the difference). These results are recorded in literature in relation to the estimated composition of the bark of conifers (^{Rowell 2013}). Regarding the infrared spectroscopy analysis, both SR and SB spectrums (Fig. 1) presented strong signal bands to OH groups intermolecular interaction (3415 cm^-1) and C-H bond stretching (2923 cm^-1). The peaks at 1510 and 1610 cm^-1 correspond to the C=C aromatic ring skeleton stretching vibration and the peak at 1054 cm^-1 corresponds to the C-O-C pyranose stretching. The peaks obtained are features to groups on carbohydrates and lignin, the major components in bract composition (^{Silverstein et al. 2005}, ^{Kumar et al. 2014}, ^{Liu et al. 2014}).

Fig. 1 Fourier transform-infrared spectra of sterile bract raw (solid line) and sterile bract boiled (dashed line) 

These signals show the presence of important polar groups in the biosorbent chemical composition. The polar groups can strongly interact by hydrogen-bonding and/or dipole-dipole interactions with compounds that possess, for example, O-H and N-H, in their chemical structures. The SR spectrum is similar to SB, indicating that the main lignocellulosic groups of biosorbent were not modified by the boiled treatment. Differential thermogravimetric analysis (DTGA) showed that SR has a lower thermal stability (182 ºC) than SB (204 ºC) (Fig. 2). This higher degradation temperature for SB can be attributed to boiled treatment that might remove or degrade the phenolic compounds. In fact, when SR is put in water, the solution becomes colored after a few minutes. This happens because its soluble phenolic compounds move to the aqueous solution. Boiling removed these compounds from SR and reduced its mass nearly by 14 %. Because of these facts, only SB has been used in rhodamine B biosorption studies.

Fig. 2 Differential thermogravimetric analysis curves of the sterile bract raw (solid line) and sterile bract boiled (dashed line) 

Rhodamine B biosorption studies

The chemical balance of biosorption between the biosorbent and the dye was reached after 2 h of contact. This time was used for all the equilibrium experiments performed in this work. The effect of pH on biosorption of rhodamine B by SB is showed in figure 3. The removal capacity of rhodamine B by SB was greater than 80 % at pH lower than 5. The minimum removal occurred at pH 10 with approximately 30 %.

Fig. 3 Effect of pH on the removal of rhodamine B by sterile bract boiled (C _i = 10 mg/L; 20 ºC) 

It is evident from the above that pH has a great influence on the biosorption behavior of naturally occurring biosorbents as well as on some adsorvates. The protonation or deprotonation of functional groups present in the biosorbent and in the dye can be pH-dependent and influence the attraction effects between charges involved in the biosorption process. For SB, the PZC was obtained at pH 6. This indicates that the biosorbent has its surface charged positively in a medium in which the pH is below 6. Above pH 6, its surface has predominantly negative electric charge. Rhodamine B is described as a cationic dye; however, the carboxyl group of its structure has pKa around 3.7 which favors interactions of charge with the biosorbent in acidic pH. For a medium in which the pH is greater than 3.7, the zwitterionic form of the dye causes the dimerization of rhodamine B. The morphology of the sterile bract has been recently described as highly porous and with a rough surface (^{Carmo et al. 2017}), and a dye dimerization can reduce biosorption by limiting its access to the sites of biosorbent interaction located within the pores (^{Huang et al. 2016}). Indeed, several studies have described rhodamine B as having a higher biosorption capacity in acid media and at pH below the PZC of the biosorbent (^{Hii et al. 2009}, ^{Dahri et al. 2016}, ^{Kooh et al. 2016}).

Based on this result, we studied the kinetics of dye biosorption by SB at pH 4 and 8. Data were applied to the kinetic models of pseudo first and pseudo second orders using non-linear regression (Fig. 4). The results confirmed a higher biosorption capacity at equilibrium for acidic pH (q _e = 0.564 mg/g) to the detriment of the alkaline medium (q _e = 0.450 mg/g). In both pH values the best data fit was obtained for the kinetic model of pseudo second order, in which the lowest χ² and ARE values were obtained (Table I). This indicated that the rate-controlling step of the biosorption process may be governed by chemical processes, which involve valency forces through the sharing or exchange of electrons between the biosorbent and adsorbate as covalent forces, and ion exchange (^{Ho 2006}).

Fig. 4 Effect of contact time and comparison of different kinetic models for rhodamine B onto sterile bract boiled (C _i = 10 mg/L; 20 ºC) in pH 4 (■) and 8 (●) (pseudo first order in dash and pseudo second order in plain) 

The equilibrium of dye biosorption by SB was studied at pH 4 and 8. Data were fitted to the Langmuir and Freundlich equations using non-linear regression (Figs. 5 and 6). For both pH values, the best fit was obtained by the Freundlich model, which achieved the smallest χ² and ARE values as well (Table II). The disagreement with the Langmuir model might occur due to the fact that the hypothesis of constant energy biosorption is not performed, because the biosorbent surface is heterogeneous. The K _F constant calculated for the experiment at pH 4 had a higher value than that obtained for the biosorption of the dye at pH 8. This confirms a higher removal capacity of rhodamine B by the biosorbent under acidic conditions. Other studies have also described the removal of dye by biosorbents as heterogeneous and a greater capacity for acid media. The removal of rhodamine B by mango endocarp (^{Inyinbor et al. 2015}) and lignin from sugarcane bagasse (^{Abou-Gamra and Medien 2013}) showed a biosorption intensity (denoted by parameter β) of a very similar value to that obtained in the present study (0.618 ± 0.0151). This parameter is also used to estimate the biosorbent surface heterogeneity. The higher the β value, the greater the surface heterogeneity (^{Vázquez et al. 2007}, ^{Lacerda et al. 2015}).

Fig. 5 Isotherm of rhodamine B biosorption on sterile bract boiled in pH 4 (C _i = 2.5 to 220 mg/L; 20 ºC). Non-linearized Langmuir isotherm (dashed line) and Freundlich (solid line) isotherm 

Fig. 6 Isotherm of rhodamine B biosorption on sterile bract boiled in pH 8 (C _i = 2.5 to 220 mg/L; 20 ºC). Non-linearized Langmuir isotherm (dashed line) and Freundlich (solid line) isotherms 

Evaluation of wastewater treatment

The evaluation of SB to treat wastewater was made using a synthetic wastewater containing commercial rhodamine B obtained from a local timber industry where the dye is dissolved in water (125 mg/L) and is stored in a big tank, which is used to dye plywood by dipping. Plywood is a sheet material manufactured from thin layers of softwood and it is dyed to identify its commercial application. The rhodamine B commercial sample was investigated by spectroscopy and showed similar characteristic bands as the analytical grade dye, which confirm its chemical identity. Treatment of synthetic wastewater containing rhodamine B within SB removed 96 % in the mass of dye (q _e = 4.95 mg/L). After this, the treatment performance was evaluated by toxicity bioassays. Untreated wastewater containing the dye caused significant growth inhibition on roots of A. cepa (Fig. 7). This phytotoxic effect could be related to bioaccumulation of rhodamine B in plants and interactions with specific areas in the meristematic cell nucleus. Due to this, different kinds of physiological, biochemical and molecular effects can be initiated, including mitotic activity change, nuclear aberration (e.g., micronucleus, nuclear buds and bridged nuclei) and oxidative stress (^{Tan et al. 2014}).

Fig. 7 Length of roots in A. cepa. ***Significant difference compared to the negative control and treated dye solution (p < 0.001)