Removal of Cu ( II ) and Zn ( II ) from water with natural adsorbents from cassava agroindustry residues

Current study employs solid residues from the processing industry of the cassava (Manihot esculenta Crantz) (bark, bagasse and bark + bagasse) as natural adsorbents for the removal of metal ions Cu(II) and Zn(II) from contaminated water. The first stage comprised surface morphological characterization (SEM), determination of functional groups (IR), point of zero charge and the composition of naturally existent minerals in the biomass. Further, tests were carried out to evaluate the sorption process by kinetic, equilibrium and thermodynamic studies. The adsorbents showed a surface with favorable adsorption characteristics, with adsorption sites possibly derived from lignin, cellulose and hemicellulose. The dynamic equilibrium time for adsorption was 60 min. Results followed pseudo-second-order, Langmuir and Dubinin-Radushkevich models, suggesting a chemisorption monolayer. The thermodynamic parameters suggested that the biosorption process of Cu and Zn was endothermic, spontaneous or independent according to conditions. Results showed that the studied materials were potential biosorbents in the decontamination of water contaminated by Cu(II) and Zn(II). Thus, the above practice complements the final stages of the cassava production chain of cassava, with a new disposal of solid residues from the cassava agroindustry activity.


Introduction
Cassava root processing industries generate solid wastes, commonly called bark (bark + inner bark) and bagasse (roughage, bran), generally disposed of in animal feed and / or as biofertilizers (EMBRAPA, 2005).
Since approximately 3 to 5% of root wastes correspond to the cassava bark (EMBRAPA, 2005), it may be estimated that Brazil produces approximately 1 million tons per year of bark.Eleven million tons per year are produced worldwide, besides the even greater amount of cassava bagasse, root processing residues and starch production.
Due to the fact that cassava crops are extant in many countries, its increasingly high production in the coming years (EMBRAPA, 2005) will generate even greater amounts of solid residues proportionally to the expansion of cultivation.Research is thus required for new types of disposals and uses for the residues.A possible solution may be the use of cassava residues as alternative biosorbents in contaminated waters.
Alternative materials such as agro-industrial residues have been evaluated due to their high availability and accessibility, efficiency and competitiveness with regard to ion exchange resins and activated carbon (VALDMAN et al., 2001).
Different plant residues have been used as biosorbents in contaminated water, such as modified sugar cane bagasse or in natura for the removal of Pb and Cu (DOS SANTOS et al., 2010;2011); banana and orange peels in the removal of Cu, Co, Ni, Zn, and Pb (ANNADURAI et al., 2002); mussel shells in the biosorption of Hg (PEÑA-RODRÍGUEZ et al., 2010); rice bran in the biosorption of Cd, Cu, Pb and Zn (MONTANHER et al., 2005); and water hyacinth dry biomass in removal of Cd, Pb, Cr, Cu, Zn and Ni (GONÇALVES Jr. et al., 2009).
Owing to the characteristics and advantages attributed to the biosorption process, assays on new adsorbent materials which are technically and economically feasible become crucial (DOS SANTOS et al., 2010).
Water pollution by heavy metals, particularly Cu in excess, may be toxic to humans, since the later metal has an affinity with SH groups of several proteins and enzymes associated with such diseases as epilepsy, melanoma and rheumatoid arthritis.Further, Cu may also lodge in the brain, liver and stomach, manifesting such symptoms as gastric ulcer, liver necrosis and kidney disease and also the loss of palatability (LIMA et al., 2006;ALMEIDA NETO et al., 2014).
In spite of its essentiality, Zn present in high concentrations may cause harmful effects on the human body, such as nausea, vomiting, diarrhea, skin irritation, anemia and infertility, affecting the respiratory system (nose to lungs), gastrointestinal tract (digestive) and blood (blood formation) (SEOLATTO et al., 2009;ATSDR, 2015).
There are some researches reporting the use of cassava solid wastes for the removal of metal ions from water: the adsorption of Cd(II), Cu(II) and Zn(II) by cassava tuber bark waste (HORSFALL Jr. et al., 2006); cassava mesocarp treated with dye for Cu(II) and Zn(II) removal (AGIRI; AKARANTA 2009); activated carbon of cassava peels for Cu(II) adsorption (MORENO-PIRAJÁN; GIRALDO 2010) and others.All the above researches report on the use of cassava bark alone (there is no mention of cassava bagasse) or the resultant adsorbent is modified by some physical/chemical process which increases the cost of the treatment and fails to be an attractive technology in undeveloped countries.
Current research employs natural organic adsorbents derived from the cassava agro-industrial residues for the removal of Cu(II) and Zn(II) from water, and assesses the kinetic, equilibrium and adsorption thermodynamic assays.

Acquirement and characterization of adsorbents
The bark´s raw material, bagasse and bark + cassava bagasse were obtained directly from an agroindustry.They were dried, milled and standardized in particle size (14 to 60 mesh), resulting in biosorbents based on bark, bagasse and bark + bagasse mixture.
The pH correspondent to the point of zero charge (pH PZC ) of the adsorbent was determined.It refers to the pH rate when the resultant of surface charges of the adsorbent is null.Further, 50 mg of adsorbents and 50 mL of aqueous solution of KCl (Vetec 99%) 0.5 mol L -1 in pHi (pH initial) rates varying from 2.0 to 9.0 (adjusted with HCl 0.1 mol L -1 and NaOH 0.1 mol L -1 ) were added.After stirring for 24 hours (200 rpm) at 25°C, pHf (pH final) rates were obtained, providing a graph of pHi versus Δ pH, with the point of null variation being pH corresponding to pH PZC (MIMURA et al., 2010).
The characterization of the infrared spectrum was performed to evaluate the possible functional groups causing the binding with Cu(II) and Zn(II), by a spectrometer Shimadzu Infrared Spectrophotometer FTIR -8300 Fourier Transform, in the region between 400 to 4000 cm -1 , with resolution 4 cm -1 .The spectrum was obtained using KBr tablets.
The surface was also evaluated morphologically by scanning electron microscopy (SEM), with a FEI Quanta 200 microscope at a voltage of 30 kV.The samples were placed in a double-sided tape attached to a carbon sample holder and then metallized with gold at a thickness of about 30 nm by a sputter coater Baltec Scutter SCD 050.
Adsorption kinetics were evaluated by measuring the amount of metal ion adsorbed through contact time between adsorbent and adsorbate (5, 10, 20, 40, 60, 80, 100, 120, 140, 160 and 180 minutes), whereas in the case of equilibrium assays the removal of metals was evaluated in solution according to the initial concentration of the metal ions (5,10,20,40,60,80,100,120,140 and 160 mg L -1 ).
The following constant conditions were employed in all the studies evaluated: 400 mg of adsorbent matter (bark, bagasse and bark + bagasse) and Cu(II) and Zn(II) solutions at pH 5.5.These conditions had been defined in previous studies (SCHWANTES et al., 2013).It should also be noted that 125 mL Erlenmeyer flasks in a Dubnoff thermostatically system with constant stirring at 200rpm were used in the adsorption studies.
After the sorption process in the Dubnoff system was performed, the samples were filtered and aliquots removed for the determination of concentrations of metals by FAAS (WELZ;SPERLING, 1999).
Rates for the equilibrium concentration were used to calculate the adsorbed amount at equilibrium (Equation 1).

=
− . ( where Q eq is the number of ions adsorbed per gram of adsorbent (mg g -1 ); m is the material of the adsorbent used (g); C 0 represents the initial concentration of ion (mg L -1 ); C eq is the ion concentration in solution at equilibrium (mg L -1 ); V is the volume of solution used (L).
The removal percentage of the metal was calculated with Equation 2: where %R is the ion removal percentage by the adsorbent; C eq is the ion concentration at equilibrium (mg L -1 ); C 0 is the initial concentration of the ion in the solution (mg L -1 ).
The results obtained in the above tests were linearized according to mathematical models of pseudo-first-order, pseudo-second-order, Elovich, intraparticle diffusion, Langmuir, Freundlich, Dubinin-Radushkevich and determination of thermodynamic parameters ΔG, ΔH and ΔS (Equations 3 and 4 R is the universal constant of gases (8.314 J mol -1 K -1 ); T is the temperature in Kelvin (K); K d is the coefficient of thermodynamic distribution, obtained by Q eq /C eq .
All evaluations of current adsorption experiment were performed in triplicate.

Results and discussion
According to Mimura et al. (2010), pH PZC may be defined as the pH of the surface of the solid with neutral charge.In solutions with pH below pH PZC , in adsorbent´s surface with predominate positive charges and in solutions with pH above pH PZC , the surface liquid charge is negative and provides a better condition for the adsorption of cations.
Results for pH PZC (Figure 1) show that the equivalence point of positive and negative charges for the adsorbents is 6.00 for bark, 6.17 for bagasse and 6.24 for bark + bagasse.Consequently, the adsorption of cations, in this case Cu(II) and Zn(II), are favored by pH rates above pH PZC [5] (TAGLIAFERRO et al., 2011).As Figures 2, 3 and 4 show, the adsorbent surface has a fibrous and porous appearance with an irregular and heterogeneous structure.The gaps indicate that the material´s characteristics suggest that the adsorbent possesses favorable conditions for the removal of ions in the solution.
The characterization by infrared (IR) contributes towards the understanding of the metal ion behavior or adsorption mechanism on a solid surface and provides information on the functional groups in the adsorbent´s structure (WAN NGAH; HANAFIAH, 2008), whereas biosorption is the result of electrostatic interaction and complex formation between the metal ions and the functional groups in the biomass (DOS SANTOS et al., 2011).Bands 3440, 2920, 1730, 1650, 1420 and 1030 cm -1 were employed for the adsorbent materials under analysis (Figure 5).
Figure 5 shows a broad and strong band at 3440-3330 cm -1 which may be attributed to the stretching vibration of the O-H bond.This band features the stretching vibration of hydroxyl groups in carbohydrates, fatty acids, proteins, lignin units, cellulose and absorbed water (FENG et al., 2011;HAN et al., 2010;STUART, 2004).Moreover, the band at 2920 cm -1 is attributed to a stretching vibration of C-H bond of the alkane groups (BARBOSA, 2007).2004), they may be attributed to the carbonyl groups of ketone and aldehyde functions in lignin and holocellulose.
The bands located between 1420 and 1650 cm -1 may be attributed to the stretching vibration of C-O bond from amide and carboxylic groups (HAN et al., 2010).The bands observed at 1030 cm -1 may be attributed to C-O stretching, suggesting the presence of lignin too (PASCOAL NETO et al., 1995), which, according to Guo et al. (2008), has compounds similar to carboxyl groups that provide adsorption sites.
Thus, according to Sharma et al. (2006) and Pehlivan ( 2009), the adsorption of metals by plant residues, natural materials and agro-industrial residues may be attributed to some functional groups such as lignin, alcohols, carboxylic groups, carbohydrates and proteins, which adsorb metal ions.
Since Pb concentrations were also detected in adsorbent materials (Table 1), this fact indicated heavy metal concentration in the soil where the cassava plants were grown.
Pb in biosorbents may have been retrieved from plants grown in soil contaminated with the heavy metal.Pb may migrate from the adsorbent biomass to the solution and contaminate the remaining solution.However, tests performed in current research have shown that Pb was strongly adsorbed by the biomass since no traces of lead concentration were detected.Gonçalves Jr. et al. ( 2011) evaluated the residual effect of fertilizers on wheat by using fertilizers from unknown origin and registered soil contamination by Pb through the fertilizers and consequent absorption of Pb by wheat plants.Similarly, it may be theorized that when grown in contaminated soils, cassava plants also absorb metals (such as Pb) in its tissues, and in the case of toxic heavy metals, they tend to accumulate them in the roots.
Table 2 still shows that the model of pseudosecond-order has better adjustment (R 2 ) and estimates satisfactorily the Q eq rates, very close to the experimental ones (Table 3).Thus, the model suggests the occurrence of chemisorption for Cu (II) and Zn (II).
As Table 2 2014) reported similar results with Loofah fibers for the removal of Cu(II) from contaminated solutions.According to their results, the model of pseudo-first-order also failed in the adsorption process, with K 1 (min. - ) = 0.0076; Q eq (mg g -1 ) = 7.33; R 2 = 0.97, which were very unlike the visualized results.
Similar results are shown by Paduraru et al. (2015) who removed Zn(II) from aqueous solutions with rapeseed waste.According to these authors, the pseudo-second-order kinetic model implies that the rate of sorption sites covering is proportional to the square of the number of unoccupied and occupied sites.
Following the results obtained, Paduraru et al. ( 2015) concluded that the sorption of Zn(II) ions on the tested rapeseed fitted better the pseudosecond-order kinetic model, with values of k 1 (g mg -1 min. - ) = 0.0022 and R 2 = 0.997, for an initial Zn(II) concentration of 48 mg L -1 .
Table 3 shows the rates derived from the linearization of the intraparticle diffusion model.
The intraparticle diffusion model explains mathematically the movement of the polluting particle from the external environment to the interior of the adsorbent pore (GUPTA and BHATTACHARYYA, 2011).However, Table 3 reveals slightly significant adjustments (R 2 ) for the model, indicating that it fails to be satisfactory to explain the sorption phenomenon.
Table 4 presents the rates obtained from the linearization of adsorption equilibrium assays.
Table 4 shows that the Langmuir model satisfactorily explains the monolayer adsorption of Cu(II) and Zn(II) by the adsorbents Bark, Bagasse and Bark + Bagasse, with maximum adsorption capacity (Q m ) for Bagasse and Bark + Bagasse for the adsorption of Cu(II) and Zn(II).
As mentioned above, the concomitant occurrence of adsorption of metal ions in mono-and multilayers may occur in the biosorption process, since the biosorbent materials are very heterogeneous and may present more or less energetic, active sites.
Gonçalves Jr. et al. ( 2012) evaluated the Pinus bark as a natural adsorbent and reported that the Freundlich model fitted best for Cd and Cr adsorption, whereas Langmuir and Freundlich models fitted best for Pb, indicating mono-and multilayer adsorption of the metal.
Since Table 4 also shows rates of E (sorption average energy) greater than 8 kj mol -1 , the occurrence of chemical adsorption between adsorbate / adsorbent was again suggested (WAN NGAH et al., 2008).
Among the mathematical models evaluated, the Langmuir and D-R models presented the best adjustments (R 2 ) and suggested the occurrence of chemisorption of Cu(II) and Zn(II) in monolayers.
The assessment of the Langmuir R L rates (Table 4) showed that the conditions under which adsorption tests were developed indicated that the adsorptive process was favorable in all cases, since rates ranged between '0' and '1' (LIN;JUANG, 2009).
Table 5 shows the results obtained in the Cu(II) adsorption thermodynamic studies.
As may be observed in Table 5, the adsorption of Cu(II) was significantly influenced by temperature increase since increase in the temperature of the environment causes a significant increase in metal removal.
Positive rates of ΔH are a strong indication of an endothermic sorption system (WAN NGAH; FATINATHAN, 2010).
According to Crini and Badot (2008), the laws of thermodynamics indicate that in a constant temperature and pressure, the ΔG value is the main criterion for the indication of system spontaneity.
Moreover, negative ΔG rates indicate the spontaneous nature of the reaction, whereas positive ΔS rates demonstrate an increase in disorder and randomness of the solid/solution interface during the sorption process, as occurred in the biosorbents under analysis (WAN; NGAH HANAFIAH, 2008).
Thus, according to rates for ΔG (Table 5), the removal of Zn(II) by the adsorbents Bark and Bagasse has a non-spontaneous and endothermic nature, while the adsorption of this metal ion by the adsorbent Bark + Bagasse becomes spontaneous at temperatures above 35°C.
Figure 6 evidences the batch temperature increase system and the consequent increase or decrease of Zn(II) adsorption.
Table 6 shows the thermodynamic parameters obtained for the removal of Cu(II) by the natural adsorbents under analysis.
In an analogous manner to the above, the ΔG and ΔH rates shown in Table 6 indicate non-spontaneous and endothermic nature for the removal of Cu(II) by the adsorbent Bark, while spontaneity of the adsorption reaction was noted for Bagasse and Bark + Bagasse

Conclusion
SEM analysis showed a fibrous and porous material with irregular and heterogeneous structure, with many gaps, suggesting that the adsorbent had favorable conditions for the adsorption of ions.IR analysis indicated bands O-H, C-H and C-O, suggesting the presence of hydroxyl, alkane and carbonyl groups from carbohydrates, proteins, lignin, cellulose and hemicellulose, which indicated the possible active sites for Cu(II) and Zn(II) adsorption Results obtained in current study demonstrated that the adsorption of Cu(II) and Zn(II) occurred in monolayers and had a chemical nature.Moreover, the above-mentioned natural adsorbents provided high removal rates for the metal ions studied.
Results show that the use of solid residues from cassava agroindustry proves to be potentially viable as biosorbents in the decontamination of water bodies polluted by metal ions.

Figure 1 .
Figure 1.Point of zero charge in KCl 0.5 mol L -1 for bark, bagasse and bark + bagasse.

Figures 2 ,
Figures 2, 3 and 4 illustrate the adsorbent micrographs at a magnification of 20,000 times.

Figure 5 .
Figure 5. Infrared spectra for the bark, bagasse and bark + bagasse.According toHorn et al. (2011), bands included in 1730 indicate the presence of starch, or, according toPastore et al. (2004), they may be attributed to the carbonyl groups of ketone and aldehyde functions in lignin and holocellulose.The bands located between 1420 and 1650 cm -1 may be attributed to the stretching vibration of C-O bond from amide and carboxylic groups(HAN et al., 2010).The bands observed at 1030 cm -1 may be attributed to C-O stretching, suggesting the presence of lignin too(PASCOAL NETO et al., 1995), which, according toGuo et al. (2008), has compounds similar to carboxyl groups that provide adsorption sites.According to Pehlivan et al. (2009), vegetal biomass primarily comprises cellulose, hemicellulose and lignin.Thus, according toSharma et al. (2006) and Pehlivan (2009), the adsorption of metals by plant residues, natural materials and agro-industrial residues may be attributed to some functional groups such as lignin, alcohols, carboxylic groups, carbohydrates and proteins, which adsorb metal ions.Since Pb concentrations were also detected in adsorbent materials (Table1), this fact indicated heavy metal concentration in the soil where the cassava plants were grown.Pb in biosorbents may have been retrieved from plants grown in soil contaminated with the heavy metal.Pb may migrate from the adsorbent biomass to the solution and contaminate the remaining solution.However, tests performed in current research have shown that Pb was strongly adsorbed by the biomass since no traces of lead concentration were detected.Gonçalves Jr. et al. (2011) evaluated the residual effect of fertilizers on wheat by using fertilizers from unknown origin and registered soil contamination by Pb through the fertilizers and consequent absorption of Pb by wheat plants.Similarly, it may be theorized that when grown in contaminated soils, cassava plants also absorb metals (such as Pb) in its tissues, and in the case of toxic heavy metals, they tend to accumulate them in the roots.Table2shows the rates for Cu(II) and Zn(II) adsorption kinetics for the adsorbents under analysis.

Table 1 .
Metal levels in adsorbent materials as bark, bagasse and bark + bagasse.

Table 2 .
Kinetic parameters obtained in the adsorption of Cu(II) and Zn(II) on bark, bagasse and bark + bagasse

Table 3 .
Kinetic parameters obtained in the study of adsorption of Cu(II) and Zn(II) on bark, bagasse and bark + bagasse by the intraparticle diffusion model and Q eq experimental rates.id : intraparticle diffusion constant; C i : suggests the thickness of the boundary layer effect; R 2 : coefficient of determination. K

Table 4 .
Parameters of the Langmuir, Freundlich and Dubinin-Radushkevich (D-R) mathematical models for the biosorption process of Cu(II) and Zn(II) on bark, bagasse and bark + bagasse.

Table 5 .
Q eq rates and the thermodynamic parameters of adsorption of Zn(II) on biosorbents bark, bagasse and bark + bagasse, C 0 = 40 mg L -1 .eq : amount of metal ion adsorbed per unit of adsorbent; ΔG: variation of Gibbs free energy; ΔH: enthalpy change; ΔS: entropy change; R 2 : coefficient of determination Q

Table 6 .
Q eq rates and thermodynamic parameters of adsorption of Cu(II) on biosorbents bark, bagasse and bark + bagasse, C 0 = 40 mg L -1 .
Q eq : amount of metal ion adsorbed per unit of adsorbent; ΔG: variation of Gibbs free energy; ΔH: enthalpy change; ΔS: entropy change; R 2 : coefficient of determination.