Adsorption of Cd2+ Ions from Aqueous Solution Using Biomasses of Theobroma cacao, Zea mays, Manihot esculenta, Dioscorea rotundata and Elaeis guineensis

In this work, the mechanisms of cadmium (Cd2+) adsorption on residual biomasses from husks of yam (Dioscorea rotundata), cassava (Manihor esculenta), cocoa (Theobroma cacao), corn (Zea mays) and oil palm bagasse (Elaeis guineensis) were studied in order to evaluate the effect of temperature, adsorbent dose and particle size in a batch system. Isotherms and adsorption kinetics were determined and adjusted to different models. The biomaterials were characterized using the techniques of Scanning Electron Microscopy (SEM) and Energy-Dispersive X-ray Spectroscopy (EDS). Results reveal that the possible mechanisms of Cd2+ adsorption in bioadsorbents were ion exchange and complexation with -COOH and -OH groups. From the experimentation, it was found that best conditions were presented at 55 °C, particle size 0.5 mm and 0.03 g adsorbent. The following biomass performance was obtained in terms of adsorption capacities: cocoa husk (CH) > corn cob residues (CCR) > cassava peel (CP) > palm bagasse (OPB) > yam peel (YP), according to the Langmuir and Dubinin- Radushkevich (D-R) models. The equilibrium of Cd2+ adsorption over YP and OPB was well described by Langmuir’s isothermal model, while for CH, CCR and CP the model that best fit experimental data was Freundlich’s model. The results of D-R model suggested that the process is controlled by physisorption mechanism with strong interactions among active sites and Cd2+ ions. The kinetics for all systems studied fit the pseudo-second order model. The values of the thermodynamic parameters established that cadmium removal is of endothermic nature and not spontaneous using YP and CP, and exothermic, spontaneous and irreversible when using OPB, CH and CCR. The results suggest the use of YP, OPB, CH, CP and CCR residues for the removal of aqueous Cd2+.


Introduction
The discharge of industrial effluents with toxic and hazardous materials is an alarming problem because of the serious water pollution that may cause [1]. Among these pollutants, heavy metals are responsible for environmental and public health issues [2]. Cadmium is a heavy transition metal that exists in the aquatic environment through geochemical processes and is increasingly exposed to the air through anthropogenic industrial activities [3], that pollute natural water sources. Principal responsible of this environmental issue are effluents from metal plating industries, Cd-Ni batteries, pesticides, paints, pigments, plastics, metallurgy, fertilizers, alloys, stabilizers and metal plating [4,5]. Cadmium is a toxic metal that, since it is not biodegradable, accumulates within living organisms. It is considered carcinogenic and inhibits the functioning of the liver, kidneys and lungs. Cadmium also causes bone degradation, cancer, high blood pressure, destruction of testicular tissue and

Characterization of Biomasses
The biomaterials were characterized by Scanning Electron Microscopy (SEM) and Energy-Dispersive X-ray Spectroscopy (EDS). It is a surface analysis technique, which consists in focusing a thin electron beam on a sample, accelerated with excitation energies from 0.1 kV to 30 kV and which allows obtaining morphological, topographical and compositional information of the samples producing high resolution images (up to 3 nm). A SEM microscope model coupled with EDS was used (JSM-6490LV, JEOL Ltd., Medellin, Colombia). The compositional analysis of the bioadsorbents was taken from previous studies in order to quantify the presence of cellulose, hemicellulose, lignin, pectin, and elements using digestion-thermogravimetry, acid digestion-thermogravimetry, photocolorimetry and AOAC 949.14 standards.

Adsorption Tests
Cadmium sulphate (CdSO 4 , analytical grade) was used to prepare the Cd 2+ stock solution at 100 mg/L [35]. A design of experiments carried out in Statgraphics Centurion XVI.II with a continuous linear factor in a central star composite response surface was applied. This design allows studying the effect of the variables that influence the response by combining them simultaneously, carrying out a limited number of experiments without the need for replications, by mapping the region of a response surface [36]. Particle size (0.14, 0.36, 0.5, 1.0 and 1.22 mm), adsorbent dose (0.03, 0.15, 0.33, 0.5 and 0.62 g) and temperature (29.8, 40, 55, 70 and 80 • C) were varied for a total of 16 experiments per bioadsorbent. The adsorbent dosage was added to 100 mL of stock Cd 2+ solution at 200 rpm and pH 6 for 24 h. The remaining concentration of Cd 2+ ions was determined using a Unicam model 969 atomic absorption spectrophotometer with acetylene/air flame. The efficiency and adsorption capacity were calculated by Equations (1) and (2), respectively: where C f is the remaining concentration of metal ions in solution is, C i is the initial concentration of metals in solution, V is the volume of solution and m is the amount of bioadsorbent used.

Kinetics and Adsorption Isotherms
Kinetics was studied under the best experimental conditions found in order to determine the equilibrium time of adsorption. For this purpose, the bioadsorbent was contacted with 100 mL of Cd 2+ solution of the same concentration already prepared in an IN-666 Shaking incubator at 200 rpm, taking aliquots at different time intervals (5,10,15,20,30,60,120,240,480, and 960 min). The experimental data were fitted to the pseudo-first order (PFO), pseudo-second order (PSO) [37] and Elovich Lagergren kinetic models to describe and understand the adsorption process.
Adsorption isotherms were performed in order to describe the equilibrium of the separated solute between the solid and liquid phase, as well as to determine the type of interactions that control the process. Different initial concentrations of Cd 2+ (25,50,75,100 and 200 mg/L) were used to perform the tests, keeping the agitation constant at 200 rpm at the best conditions found during the adsorption tests. The experimental data were adjusted to Langmuir, Freundlich and Dubinin-Radushkevich's models [38].

Adsorption Thermodynamics
Calculation of thermodynamic parameters will be done by the graphical method based on the Van 't Hoff equation, in order to determine spontaneity, feasibility, type of adsorption and to predict the magnitude of changes on the surface of the adsorbent respectively, as well as the effect of temperature on the process. The change in the Gibbs standard free energy (∆G • ), standard enthalpy (∆H • ) and standard entropy (∆S • ) was estimated by testing the adsorptive capacity of the samples by varying the temperature from 30 to 80 • C. Moreover, the final metal ion concentration after adsorption was determined and the amount of metal ion removed was estimated. The calculation of the thermodynamic parameters was carried out using the equations below [39].
where, K c is the equilibrium constant, q e is the concentration in the solid phase at equilibrium (mg/g), and C e is the concentration in solution at equilibrium (mg/g). The R is the constant of the ideal gases 8314 J/mol*K, T is the absolute temperature at Kelvin. The change in standard enthalpy (∆H • ) and entropy (∆S • ) are determined from the slope and the y-axis intercept of lnK c vs. T −1 , respectively.

Characterization of Bioadsorbents
African palm fibrillates (OPB) were obtained from the Department of Bolivar-Colombia, as a result of the waste generation during palm oil extraction process. The hulls of hawthorn yam (YP), cassava (CP), cocoa (CH) and corn residues (CCR). The morphological and compositional characteristics of these materials were analyzed. Figures 1-5 are referring to biomaterials before and after adsorption characterized via SEM analysis coupled with EDS. The heterogeneous surface of the biomaterials can be clearly seen. Most of the biomasses showed a smooth porous structure. According to Warui at al. [40] molecules in the amorphous region contain hydroxyl groups specially for cassava peel. In SEM micrographs before and after adsorption, luminous and non-luminous features are observed, which revealed the presence of organic minerals such as Ca 2+ , Mg 2+ and K + in the matrix of biomasses [41]. The EDS spectra also confirmed these findings with characteristic peaks attributed to these elements. Before adsorption, no signal was recorded for cadmium ions.          As shown in Figures 1a and 3a, the elements C, O, Ca, K and Si were present in greater proportion in the structure of biomaterials, with YP and CH being the most diverse in structure. The higher presence of these mineral parts (revealed by small white spot) is also confirmed by the elemental compositional analysis gathered from the literature. After adsorption, the decrease in K near the 3.3 keV peak, the reduction of Ca by 1.3 keV and its disappearance by 3.4 keV, as well as the disappearance of Fe are observed in the YP; OPB shows a decrease of Ca in 1.3 keV and 3.4 keV and of Si in 1.6 keV; the CH had a decrease of K in 3.3 keV; CCR showed disappearance of Si at 1.9 keV and of K at 3.3 keV; On the other hand, an increase in K was observed in the CP spectrum with the appearance of Cd near 2.8 keV. The aforementioned occurred together with the conservation in most cases of the presence of Ca in the 1.3 keV peak, finding conservation of the variation of the equivalent charges between Cd 2+ and Ca 2+ /K + , hence, adsorption of Cd 2+ is more competitive for functional groups that resulted in the ion exchange process [42]. The increase of the Si atomic percentage in OPB; P, O and S in CH; Si, P and O in CP; and P, Mg, Al, O and K in CCR, after Cd 2+ adsorption illustrates that the formation of precipitation in the form of calcium sulfate and phosphate on the cell surface, which was observed in SEM micrographs, was another biosorption mechanism probable. Table 1 shows the chemical composition of the bioadsorbents under study, with carbon being the most abundant element in all bioadsorbents, which is typical of lignocellulosic materials due to the presence of lignin, cellulose and hemicellulose, which is why expects a high Cadmium adsorption efficiency, because these polymers are known for the large number of hydroxyl and carboxylic groups that can favor the adsorption of heavy metals [3,15]. Previous research reports that the presence of OH, COOH and amines groups contribute to the metal ion uptake process in the lignocellulosic matrix of bioadsorbents, which favors interaction with the active centers due to their anionic nature [43,44]; In the adsorption of Cd 2+ on cocoa mesocarp, it was found that the functional groups involved in the removal process were those that have an oxygen in their structure such as ethers (aromatic, olefinic or aliphatic), esters, amines, hydroxyl, and aliphatic hydrocarbons [45].
Likewise, it is observed that the highest lignin content occurs in YP and OPB, therefore the removal efficiency may be higher in these materials [46]. The mechanisms of cadmium adsorption onto biomasses can include ion exchange, micro-precipitation, complexation and coordination owing to the presence of functional groups such as hydroxyl, carboxyl, amides and phenols in the lignocellulosic materials [7]; however, these work identified the cation exchange mechanism and the microprecipitation mechanism as crucial in the biosorption of Cd 2+ by the evaluated biosorbents [47,48].

Adsorption Capacity of Bioadsorbents
In the experimentation carried out to evaluate the adsorption capacity of Cd 2+ ions using the biosorbents prepared OPB, CCR, YP, CP and CH, the effect of temperature, particle size and adsorbent dose in the process was studied. The adsorption capacity of the metal was calculated using Equation (2). As summarized in Table 2, the highest absorption capacity of Cd 2+ is obtained for all the biosorbents YP, OPB, CH, CP, CCR, at the conditions of temperature 55 • C, particle size 0.5 mm and dose of 0.03 g adsorbent, obtaining values of 123. 66,197.95, 160.83, 76.86, 105.40, mg/g respectively. It can also be observed that the process is endothermic, because the behavior of the Cd 2+ adsorption capacity of all the biomaterials studied shows that an increase in temperature favors removal, in the range of 40 • C to 80 • C. The increase in the removal efficiency of Cd 2 + due to the increase of this parameter could be due to the union between the adsorbent and adsorbent by the formation of new active sites, these sites could form new bonds between ions and the active functional groups in the adsorbent, overcoming the activation energy barrier and improving the intraparticle diffusion rate bonding [5,53,54]. The reduction in adsorption capacities at temperature of 70 • C for the CH, CP y CCR is attributed to the tendency of the process to desorb when adding higher heat not needed, more cadmium ions escape from the solid phase [53,55]; from elsewhere, YO y OPB show that the adsorption capacity increases as a function of temperature at the different conditions of particle size and dose of adsorbent evaluated.
It is observed that the adsorption capacity behaves inversely proportional to the increase in particle size of YP and OPB, which may be due to the fact that a decrease in particle size the surface area is larger due to the number of pores, thus increasing the active junction sites and the available contact surface, resulting in a higher mass transfer and faster adsorption [56,57]. On the other hand, when using CCR, CH and CP, it was found that 0.5 mm particle diameter offers the best results, which is due to the interactions of sorbate-sorbent on the surface. This behavior can be attributed to the relationship between the effective specific surface area of the adsorbent particles and the size of the particles [58,59]. Increasing the amount of the evaluated adsorbents in contact with the Cd 2+ ions contaminated solution has a negative effect on the adsorption capacity when particle sizes are larger than 0.5 when using OPB, CCR, YP, CH and CP, which may be due to the decrease of the surface area available for ion exchange due to the increase in particle size [60]. This behavior is due to the high availability of active sites in bioadsorbents because of the presence of the hydroxyl, carbonyl, amine and unsaturated hydrocarbon groups characteristic of lignocellulosic materials [61][62][63].
These findings were compared with previous publications for the removal of heavy metal ions using biomasses such as the selected in this work. Table 3 summarizes the selected works and the maximum adsorption capacities reached at defined conditions. The capacities reached in this work for the selected biomasses are within the range observed for lignocellulosic biomasses under similar conditions of temperature and pH. These results are for an adsorption cycle, however, high desorption capacity is reported after up to fourth cycles using different desorbing agents such as HCl, NaOH and HNO 3 , since it was possible to recover 52.47 and 74.84% of the metal for nickel and lead, respectively [33]. It is expected that for cadmium, being a cation with a +2 charge, a similar behavior will be obtained with the same biomaterials. Figure 6 shows the fit of the kinetic data to the PFO, PSO, Elovich and intraparticle diffusion models, in order to identify the stages that control the process. It was found at best experimental conditions presented at 55 • C, particle size 0.5 mm and 0.03 g adsorbent, the saturation time of the bioadsorbents evaluated is at 60 min and that equilibrium is reached after 120 min, indicating that the availability of active sites in the biomass is reduced due to the occupation of these by the metal ion [67][68][69]. The values of the fitting parameters of the evaluated models and their R 2 are recorded in Table 4. Taking into account the calculated parameters, it was found that the PSO and Elovich models are the best fitting models for the experimental Cd 2+ adsorption data on YP and CP, this is also observed in Figure 6a,d, According to this result, the process happens by chemisorption in the system, in which the adsorption rate is limited by valence forces given by the exchange of electrons between the adsorbate and the adsorbent due to the non-homogeneous surface of the solid [70][71][72].     Figure 6b depicts that the ion removal data on OPB biomass present a good fit to all the models suggesting that the rate of adsorption depends on a mechanism that acts on the active sites on the surface of the biomass by chemisorption [57]. Adjusting the Cd 2+ adsorption kinetics on OPB suggests that adsorption is carried out in two steps: the first attributed to the diffusion of metal through the solution to the external surface of the adsorbent, which would explain the rapid rate of adsorption during first minutes [73,74], and the second indicates the diffusion of the intraparticle metal into the pores of the OPB. The second step did not go through the origin, indicating that the dominant mechanisms during the process were intraparticle diffusion and chemical adsorption [75]. From Figure 6c,e, there are not significant adjustment between models and experimental data for CH and CCR suggesting that these biomasses are not well governed by the principles supported in such models.

Adsorption Kinetics
According to the adsorption capacity of the biomass in equilibrium, it is established that the selectivity of the adsorbents for the Cd +2 ion obeys the following order: OPB > YP > CP > CCR > CH; This biomass behavior could be explained by the content of lignocellulosic material within its matrix (Table 1), which is intrinsically linked to the presence of functional groups involved in adsorption such as OH, COOH, amino, among others [7]. It was found that the biomaterials with the highest lignin content were YP and OPB, hence, the removal efficiency was higher in both adsorbents [46].

Adsorption Isotherms
The effect of the initial concentration on the removal of Cd 2+ was evaluated by varying the initial concentration in intervals of 25 and 100 mg/L at pH 6 in order to understand the driving forces behind the adsorption process. Figure 7 shows the adjustment of the experimental data to Langmuir, Freundlich and Dubinin-Radushkevich's models after 24 h of contact in a batch system; the adjustment parameters of the models are shown in Table 5.
The R 2 parameter close to 1 for all the biomasses in Langmuir and Freundlich models suggested that the materials are well described by these isotherms; however, the sum of squares provides additional information to consider. For YP and OBP biomasses, the SS was lower for Langmuir model, while for CH, CP and CCR biomasses, the Freundlich model best fits the data. Hence, the YP and OBP follows a monolayer adsorption and the CH, CP y CCR a multilayer process [71].
ing the initial concentration in intervals of 25 and 100 mg/L at pH 6 in order to understand the driving forces behind the adsorption process. Figure 7 shows the adjustment of the experimental data to Langmuir, Freundlich and Dubinin-Radushkevich's models after 24 h of contact in a batch system; the adjustment parameters of the models are shown in Table 5. The R 2 parameter close to 1 for all the biomasses in Langmuir and Freundlich models suggested that the materials are well described by these isotherms; however, the sum of squares provides additional information to consider. For YP and OBP biomasses, the SS was lower for Langmuir model, while for CH, CP and CCR biomasses, the Freundlich model best fits the data. Hence, the YP and OBP follows a monolayer adsorption and the CH, CP y CCR a multilayer process [71].
According to the values of n, the adsorption process on Cd 2+ is favorable on YP, CH, CP and CCR, and that bioadsorbents have a good affinity for ion removal [5,76]. The values lower than 8 kJ/mol reported for the mean energy of adsorption of ions by sorbate (E) calculated for the Dubinin-Radushkevich model suggest that the process is controlled mostly by the fission absorption mechanism with strong interactions between active centers and Cd 2+ ions [77]; This model presents a good fit, since the calculated maximum adsorption capacity (qDR) values are the closest to the experimental ones with R 2 ≥ 0.The above assumes that the bioadsorbents under study have a heterogeneous structure, which is required by the Dubinin-Radushkevich isotherm [78]. According to the values of n, the adsorption process on Cd 2+ is favorable on YP, CH, CP and CCR, and that bioadsorbents have a good affinity for ion removal [5,76]. The values lower than 8 kJ/mol reported for the mean energy of adsorption of ions by sorbate (E) calculated for the Dubinin-Radushkevich model suggest that the process is controlled mostly by the fission absorption mechanism with strong interactions between active centers and Cd 2+ ions [77]; This model presents a good fit, since the calculated maximum adsorption capacity (q DR ) values are the closest to the experimental ones with R 2 ≥ 0.The above assumes that the bioadsorbents under study have a heterogeneous structure, which is required by the Dubinin-Radushkevich isotherm [78].

Thermodynamic Parameters
To study the nature of the adsorption process of Cd 2+ on YP, OPB, CH, CP and CCR, the free Gibbs energy ∆G • (KJ/mol), enthalpy ∆H • (KJ*mol −1 *K −1 ) and entropy ∆S • (KJ/mol) are determined according to Equations (3)- (5). The results of the thermodynamic adsorption parameters of Cd 2+ are shown in Table 6. According to these thermodynamic findings, the positive value of ∆H • when using YP and CP indicates that the process of removing Cd 2+ on these bioadsorbents is endothermic and energy must be supplied to the system for the process to occur [5,79]; however the ions adsorb best at intermediate and low temperature values as shown by the adsorption results. The negative values of enthalpy when using OPB, CH and CCR suggest that the adsorption process is exothermic and the system releases energy, which means energy savings when scaling up the process [39,80]. Negative values of ∆S • when using YP, OPB and CCR suggest that the Cd 2+ biomass link is strong, the biosorbents have high affinity and selectivity for the metal ion studied and the randomness at the interface is lower, thus obtaining a low possibility of reversibility, which implies that the adsorption process is energetically stable [38,53].
At all temperatures evaluated, it was obtained that ∆G • is negative when using OPB and CP, which indicates that the adsorption process is spontaneous, feasible and favorable; the change in Gibbs' free energy increases with temperature indicating that spontaneity decreases proportionally with the increase in temperature [39,79,80]. On the other hand, by using YP, CH and CCR, ∆G • is positive and increases with temperature, so the process loses spontaneity and is not of a favorable nature [81]; it is also observed that as the temperature increases the ∆G • increases in magnitude, which reveals that the system evolved by itself, becoming spontaneous and favorable naturally, which would favor the removal [51].
These promising findings lead to new research routes towards the design of composite materials [82], modification of the current kinetic and isotherm models [83], and heavy metal detection studies using bio-indicators associated with micro-organisms and phytoindicators [84,85], which enables monitoring the presence of heavy metals in surface water bodies.

Conclusions
The present study concluded that: (i) That the biosorbents YP, OPB, CH, CP, CCR are potentially good for adsorbing Cd 2+ , at the optimal experimental conditions found: temperature of 55 • C, particle size 0.5 mm and quantity of adsorbent 0.03 g, and that the variable with the greatest influence on the process was the dose of adsorbent, achieving adsorption capacities between 76 and 197 mg/g. (ii) The SEM-EDS analysis suggests that the Cd 2 + adsorption mechanisms were ion exchange and complexation with the -COOH and -OH groups present in the structure of the cellulose, lignin, pectin and hemicellulose molecules of the biomasses studied. (iii) The adsorption kinetics of Cd 2 + was fast in the initial minutes due to the availability of active centers, reaching equilibrium at 60 min, being the PSO and Elovich models the ones that best fit the experimental data, suggesting that the process occurs by chemical reaction. (iv) The adsorption isotherms using YP and OPB were adjusted using the Langmuir model, indicating that the metal is adsorbed on the metal surface in a monolayer; while the balance of adsorption in CH, CCR and CP is adjusted by means of the Freundlich model, establishing that the process occurs by chemical reaction in multilayers. The Dubinin-Radushkevich model suggests that the Cd 2 + ion adsorption process is controlled by the fission absorption mechanism. (v) From the calculation of the thermodynamic parameters it was shown that the process is endothermic and not spontaneous using YP and CP, and that it is exothermic, spontaneous and irreversible when OPB, CH and CCR are used. Based on these findings, the biomasses studies reached promising adsorption capacities in the removal of cadmium ions; however, the materials from oil palm bagasse and cocoa husks received showed remarkable performance compared to the peels and corn residues.