Ecofriendly Biosorbents Produced from Cassava Solid Wastes: Sustainable Technology for the Removal of Cd 2+,Pb 2+,and Cr total

Abstract

This research is aimed at investigating the possible use of cassava agroindustry solid wastes in manufacturing adsorbents and their use in removing heavy metals Cd²⁺, Pb²⁺, and Cr^total from water. Thus, a pilot study was conducted in two main steps: (1) obtaining and characterizing the adsorbents and (2) laboratory studies focused on the evaluation of critical physicochemical parameters on adsorption, such as pH of the solution containing heavy metals, the effect of adsorbent dose, besides kinetics and equilibrium adsorption and desorption studies. Three adsorbents were studied, cassava barks, bagasse, and their mixture. SEM, FTIR, pH_PZC, acid digestion, and chemical composition analysis were employed for adsorbent characterization. The pH of the contaminated solution was evaluated within 4.0 to 7.0, while the adsorbent doses varied from 5.0 to 24.0 g L^-1. The adsorption kinetics was evaluated within 5 to 180 minutes and interpreted using pseudofirst- and second-order models. Finally, equilibrium and desorption studies were performed by evaluating adsorbent performance within 5 to 200 mg L^-1 of heavy metals, using several nonlinear models for results interpretation. SEM analysis reveals a heterogeneous structure full of cavities. FTIR before and after adsorption reveals gaps related to missing functional groups, suggesting a significant role of alkenes, carboxylic acid, alcohol, anhydride, and ether. pH_PZC is found at pH 6.02, 6.04, and 6.26 for adsorbents derived from barks, bagasse, and their mixture. In low concentrations of metals, the higher adsorption capacities were found at pH 7.0 (94.9%) using 16 g L^-1 of adsorbent, with the most cost-benefit dose found using 8.0 g L^-1. The removal of metals reaches equilibrium within 5-10 minutes of contact time with pseudosecond-order best adjustments to the observed phenomena. The adsorption of metals by a cassava adsorbent is better adjusted to the Freundlich model, with significant and critical information provided by Sips, Redlich-Peterson, Temkin, Liu, and Khan models. Adsorption/desorption studies indicate that cassava adsorbent performs, on average, -10% of the adsorption of metals compared to activated carbon. Nevertheless, factors such as low cost and availability favor the use of such natural materials.

1. Introduction

In the last decade, the increasing industrialization and urbanization have raised the number of pollutants disposed of in the environment, especially water bodies [1]. In this sense, among the many contaminants that affect aquatic environments and human health, the pollution from organic or inorganic compounds, such as toxic metals and metalloids (Cd, Pb, Cr, As, Hg, among others) and pesticides (organophosphates, carbamates, triazines, organochlorines, among others.) need to be treated as a serious issue by governs and institutions. Moreover, access to fresh water and sanitation still is a global concern. For example, WHO [2] reports that 785 million people lack even a basic drinking water service, including 144 million people who are dependent on surface water; globally, at least 2 billion people use a drinking water source contaminated with feces, that is not to mention water pollution by chemicals, such as toxic heavy metals.

Developing and undeveloped countries already struggle with the standard procedures and cannot afford more advanced technologies to treat emergent pollutants such as pesticides and metals [3]. Among the advanced treatments that are usually employed for the removal of toxic metals, we can highlight chemical and physical precipitation, ionic exchange, Fenton and photo-Fenton processes, oxidation processes, filtration by membranes, osmosis processes, advanced oxidation processes, and adsorption with activated carbons; this last example is considered a gold standard for the removal of many organic and inorganic pollutants from waters and wastewaters [4], especially when those are in low (or trace) concentrations but not low enough to ensure the safety of the disposal of the wastewater or the use of water [5–7]. That is the case with heavy toxic metals [8].

The current legislation for drinking water by Environmental Protection Agency [9], World Health Organization [4], and Ministerio de Salud de Chile [10], respectively, state the maximum contaminant level for the studied metals: Cd²⁺ (0.005 mg L^-1, 0.003 mg L^-1, and 0.01 mg L^-1), Pb²⁺ (0.015 mg L^-1, 0.01 mg L^-1, and 0.05 mg L^-1), and Cr^total (0.01 mg L^-1, 0.05 mg L^-1, and 0.05 mg L^-1). Thus, the limits are low because such substances can promote damage to living organisms even at trace levels. Further information regarding toxicity and environmental issues caused by heavy metal exposure (with special attention to Cd, Pb, and Cr) are shown in the Supplementary Materials.

Another problem that causes significant concern in the agroindustry business is the correct biomass solid waste destination, e.g., solid byproducts from the cassava industry. These agroindustries tend to generate significant amounts of organic wastes, which can or cannot be correctly destined [3, 11–13], usually disposed of either by burning, dumping, or unplanned landfilling [14, 15].

In cassava agroindustry, solid wastes are generated in the processing of cassava roots, mainly used to manufacture animal feed and biofertilizers. Schwantes et al. [13] estimate that cassava peel makes up approximately 3 to 5% of the total mass of roots and about 1 million tons of cassava peels are produced annually in Brazil, 11 million tons worldwide. This estimate does not consider the produced bagasse, representing more than 80% of the cassava industry’s solid wastes.

The last reports evidence the possibility of using cassava as adsorbent for the removal of Cd²⁺, Pb²⁺, and Cr^total from waters [16], or as biomass modified with H₂O₂, H₂SO₄, or NaOH 0.1 M [13], or for the removal of Cu²⁺ [17], Cu²⁺, and Zn²⁺ [18]. Rajeshwarisivaraj et al. [19] report the use of cassava wastes impregnated with H₃PO₄ in activated carbon development to remove dyes and metal ions. Thompson et al. [20] reported using activated biomass from cassava wastes to remove Pb²⁺ from water. Although the above reports are interesting, Schwantes et al. [18], Schwantes et al. [13], and Schwantes et al. [16] present only preliminary studies, with many gaps regarding the adsorption process and its evaluation. Also, the above authors do not elucidate the adsorption mechanism that rules metals’ removal. Jorgetto et al. [17] only used the minced cassava root as an adsorbent raw material, not focusing on the cassava bagasse, which corresponds to more than 90% of the wastes of cassava agroindustry. Moreover, the above authors only test the adsorbent for Cu²⁺ adsorption without the necessary deep required for adsorption studies (kinetics, equilibrium, and thermodynamic studies). Rajeshwarisivaraj et al. [19] tested cassava wastes as adsorbents for the removal of dyes (Rhodamine-B, direct brown, procion orange, acid violet, malachite green, and methylene blue) and metals (Cr⁶⁺, Hg²⁺, and Fe²⁺) without evaluating the influence of essential parameters, such as pH, adsorbent doses, contact time, equilibrium, and thermodynamic studies. Thompson et al. [20] developed activated carbon from solid cassava wastes, resulting in a material gifted with 119.6 m² and good adsorption capacity, nevertheless, with inferior results comparing other adsorbents and activated carbons derived from adsorbent materials. Therefore, there is still a lack of studies regarding cassava solid wastes for adsorption purposes. Thus, such raw materials are still undervalued and unexploited.

In this scenario, this research attempts to propose solutions for two problems (environmental and public health) at the same time: first, to develop natural ecofriendly adsorbents from cassava agroindustry solid wastes for the remediation of water containing toxic metals (Cd²⁺, Pb²⁺, and Cr^total), providing an economical and novel solution for water and wastewater treatment, especially attractive for undeveloped/developing countries; and the second part of the double solution, to provide a correct and possible destination for the risen tons of agro-industrial cassava wastes, which already are, in some cases, environmental problems in developing countries located in South America, Africa, and equatorial parts of Asia.

Thus, until this date, the potential of solid cassava wastes has not been fully accessed in developing ecofriendly and efficient adsorbents, especially for removing heavy metals such Cd²⁺, Pb²⁺, and Cr^total. As a consequence, this research aimed at (i) developing natural adsorbents from cassava agroindustries and studying its morphology and other essential characteristics; (ii) evaluating these natural materials regarding their adsorption capacity and performance; (iii) studying the adsorption mechanism of Cd²⁺, Pb²⁺, and Cr^total, evaluating the interactions between the pH of the metal solutions versus dose of adsorbents, as well as studies of kinetics and equilibrium; and (iv) evaluating the possibility of the cassava adsorbents reuse in new cycles (desorption studies).

2. Material and Methods

2.1. Obtaining, Development, and Characterization of the Cassava Natural Adsorbents

In agroindustry, the cassava wastes were obtained from Toledo, Paraná State (Brazil). During the cassava root preparation for further processing, the first solid waste is obtained (cassava barks) by the washing and mechanical peeling of the roots, constituted mainly of lignin and cellulose. After removing starch and other essential components, the second solid waste (bagasse) was obtained at the end of the industrial process. The bagasse is constituted mainly by lignin, cellulose, and residues of nonextracted starch.

It is essential to mention that the agroindustrial cassava sector destines these wastes for animal feeding by mixing a small proportion of barks with bagasse selling them to cattle and cow farmers as a low-cost feed. Therefore, these wastes were sampled and sent to the laboratory, washed with water, dried, milled, and sieved, with particle size standardized between 14 to 60 mesh, as already well reported by previous studies [18]. This procedure generated the three adsorbents of cassava: CASS-BK (adsorbent of cassava barks), CASS-BA (adsorbent of cassava bagasse), and CASS-BB (the mix of barks and bagasse of cassava). These three were characterized by studying their surface by pH_PZC analysis, micrographs by SEM, surface functional groups by FTIR, and their chemical composition.

The pH corresponding to the point of zero charge (pH_PZC) is defined as the pH at which the surface of the solid has a neutral charge. The methodology used in this work for its determination is called the “11 point experiment” [3]. The procedure consisted of mixing 25 mg of the adsorbent in 50 mL of aqueous solutions of KCl 0.5 and 0.05 M under 11 different conditions of initial pH, varying from 2.0 to 13.0 by adjustment with HCl and NaOH (0.1 mol L^-1), and measuring the pH after 24 h of equilibrium. The results are expressed through the graph of final pH versus initial pH, with the pH_PZC corresponding to the range in which the final pH remains constant (regardless of the initial pH), i.e., the surface behaves like a buffer. The surface micrographs were obtained using scanning electron microscopy (SEM - JEO JSM 6360-LV, Thermo®, U.S.A), equipped with dispersive energy microscopy. The surface functional groups were characterized using an FTIR-8300 Fourier transform spectrometer operating in transmission mode (4000–400 cm⁻¹), using the standard KBr pellet method (Shimadzu®, Japan). The chemical composition of metals was analyzed by acid digestion of the cassava adsorbents [70], followed by FAAS determination. Thus, 0.5 mg of each adsorbent was weighted in test tubes with 5 mL of HNO₃+HClO mixture (2 : 1 $v/v$ ), maintained at 160°C for two hours and then at 260°C for another 1.5 h. After that, samples were transferred to 50 mL volumetric flasks. The concentration of K (wavelength of 766.5 nm), Ca (422.7 nm), Mg (285.2 nm), Fe (248.3 nm), Cu (324.8 nm), Zn (213.9 nm), Cd (228.8 nm), Pb (217.0 nm), and Cr (357.9 nm) was determined by Flame Atomic Absorption Spectrometry (FAAS, GBC 932 AA) [21].

2.2. Preparation of the Metallic Solutions

All single-element solutions of Cd, Pb, and Cr were prepared by using salts of cadmium (Cd(NO₃)₂ 4H₂O $\text{PA}\ge 99.0$ % Sigma-Aldrich), lead (Pb(NO₃)₂ $\text{PA}\ge 99$ %, Sigma-Aldrich), and chromium (Cr(NO₃)₃ 9H₂O $\text{PA}\ge 99$ %, Sigma-Aldrich).

2.3. Adsorption Studies (Cd²⁺, Pb²⁺, and Cr^total)

2.3.1. pH and Adsorbent Dose Studies

The effect of the pH of the single-element solutions of Cd²⁺, Pb²⁺, and Cr^total and its interaction with different cassava adsorbent doses was evaluated by testing three (3) pH conditions (4.0, 5.0, and 6.0, adjusted with NaOH and HCl 0.1 M) and six (6) adsorbent doses (4, 8, 12, 16, 20, and 24 g L^-1). Thus, in Erlenmeyer’s of 125 mL, 200, 400, 600, 800, 1000, and 1200 mg of CASS-BK, CASS-BA, and CASS-BB were set in contact with 50 mL of the single-element solutions of Cd²⁺, Pb²⁺, and Cr^total (10 mg L^-1). The experimental condition of this preliminary study was 200 rpm at 25°C. At the end of 1.5 h of stirring, samples were taken, and the remaining metal concentration was determined by FAAS.

Using Statistica® 7.0, a response surface analysis (RSA) was conducted to interpret these preliminary results. Therefore, a weighted linear regression was used to study the possible interactions between the adsorbent doses and the solution pH. In addition, the Fisher test was conducted to determine the statistical significance of the variables above on the removal % of heavy metals. Finally, the Tukey test evaluated the removal percentage means at 5% of probability to create a ranking between the adsorbents and the evaluated metals. The studied pH range is explained by the Theoretical speciation of Cd²⁺, Pb²⁺, and Cr^total (Puigdomenech, 2018), presented in Figure S2 [22].

2.3.2. Kinetics Studies

In order to evaluate the effect of the contact time between the cassava adsorbents and the Cd²⁺, Pb²⁺ and Cr^total single-element solutions, the adsorbent dose of 8.0 g L^-1 (previous best condition) was tested in the following contact time ranges: 5, 10, 20, 40, 60, 80, 100, 120, 140, and 160 min.

Thus, in Erlenmeyer of 125 mL, 400 mg of CASS-BK, CASS-BA, and CASS-BB was set in contact with 50 mL of single-element solutions of Cd²⁺, Pb²⁺, and Cr^total solutions (10 mg L^-1), and samples were taken at each period (5 to 160 min), centrifuged at 2000 rpm for liquid-solid separation, and the remaining metal concentration was determined by FAAS.

To elucidate the kinetics of adsorption for Cd²⁺, Pb²⁺, and Cr^total into the CASS-BK, CASS-BA, and CASS-BB surfaces, the obtained data were studied by the models of pseudofirst-order (PFO) [23] and pseudosecond-order (PSO) [24] models, as exhibited in Table S1 (supplementary material).

2.3.3. Equilibrium Studies and Side-by-Side Comparison (Activated Carbon)

In order to construct isotherms for CASS-BK, CASS-BA, and CASS-BB for the adsorption of metals, in Erlenmeyer of 125 mL, the dose of 8.0 g L^-1 (400 mg) of cassava adsorbents was set in contact with 50 mL in increasing concentrations of Cd²⁺, Pb²⁺, and Cr^total single-element solutions (5, 20, 40, 60, 80, 100, 120, 140, 160, and 200 mg L^-1). The following experimental conditions were set at pH 5.5, 60 min of stirring, 200 rpm, and 25°C. The content (adsorbent solution) was centrifuged at 2000 rpm for liquid-solid separation and sampled to determine the residual metal concentration by FAAS.

It is also essential to highlight that for the isotherm construction and evaluation, a side-by-side comparison was performed with a gold standard (commercial activated carbon Synth) in the same physical-chemical conditions CASS-BK, CASS-BA, and CASS-BB.

For the interpretation of the obtained data, the nonlinear models of Langmuir [25], Freundlich [26], Sips [27], Temkin and Pyzhev [28], Redlich-Peterson [29], Toth [30], Liu et al. [31], Khan et al. [32], and Khan et al. [33] were employed. These and other equations used in this research are exhibited in Table S1 (supplementary material).

2.3.4. Desorption studies

The possibility of reusing CASS-BK, CASS-BA, and CASS-BB was evaluated through an acid elution using HCl 0.1 mol L^-1. For that, after the isotherm’s construction (after one adsorption cycle in different metal concentrations), the cassava adsorbents were dried at $60\pm 2$ °C for 24 h, and their mass was set into contact with 50 mL of HCl 0.1 mol L^-1. Then, after 60 min of 200 rpm stirring, samples were taken and centrifuged at 2000 rpm to separate the adsorbent acid solution, and FAAS determined the desorbed concentration of metal (Cd²⁺, Pb²⁺, and Cr^total). The estimative of the desorption rate ( $D$ ) is exhibited in Table S1.

2.3.5. The Effect of Temperature on the Adsorption of Cd²⁺, Pb²⁺, and Cr^total by Cassava Adsorbents

To evaluate the effect of temperature on the adsorption of heavy metals, 400 mg of each adsorbent was weighed in 125 mL Erlenmeyer flasks and set in contact with 50 mL of Cd²⁺ (50 mg L^-1), Pb²⁺ (100 mg L^-1), or Cr^total (50 mg L^-1) solutions at pH 5.5. These flasks were stirred for 60 min at 25, 35, 45, 55, and 65°C. After that period, samples were taken and centrifuged at 2000 rpm to separate the adsorbent from the solution, and FAAS determined the remaining concentration of each metal. Thermodynamic parameters ∆G°, ∆H°, and ∆S°, were estimated according to Table S1.

2.4. Data Analysis

Maple 13®, Statistica 7®, Sisvar 5.6 [34], and Origin 2018® were employed to evaluate the aforementioned nonlinear models, besides the construction of the graphics. All values from adsorption studies were obtained in triplicate. Also, the studied models were subjected to reduced chi², $R$ -square (COD), and adjusted $R^2$ analysis.

3. Results and Discussion

3.1. Characterization of the Cassava Natural Adsorbents

The CASS-BK, CASS-BA, and CASS-BB surfaces revealed by SEM (Figure 1) reveal adsorbents endowed with fibrous and spongy aspects, with irregular and heterogeneous structures. In addition, it is possible to observe some cavities, which could indicate the structure’s porosity, which is suggested as a promising adsorbent characteristic by Conradi Jr. et al. [3].

Figure 1

SEM of CASS-BK ((a, b) 800 and 6000x), CASS-BA ((c, d) 200 and 6000x), and CASS-BB ((e, f) 1200 and 5000x).

(b) (c) (d) (e) (f)

Figure 2 and Table S2 show infrared spectroscopy absorption by frequency regions from before and after the adsorption of Cd²⁺, Pb²⁺, and Cr^total. The abovementioned results highlight the missing peaks at 760 cm^-1 for CASS-BK adsorption of Cd²⁺ and Pb²⁺, suggesting C=C medium to strong bending of alkenes [35].

Figure 2

FTIR obtained for the cassava natural adsorbents CASS-BK (a), CASS-BA (b), and CASS-BB (c) before and after the adsorption of Cd²⁺, Pb²⁺, and Cr^total. Additional information is provided in Table S2.

(b) (c)

Comparing the results from before and after adsorption, there are many gaps for CASS-BA, such as 769, 1030, 1157, 1240-1662, and 2930 cm^-1 for Cd²⁺ adsorption, suggesting a significant role by C=C from alkenes; CO-O-CO from anhydride; C-O from alcohol and aliphatic ether; C-H and O-H from alkanes, carboxylic acid, and alcohol; and C-H stretching from alkanes [36]. For Pb²⁺ adsorption by CASS-BA, gaps were found at 770, 1039, and 1155 cm^-1, indicating C=C bending from alkenes, CO-O-CO stretching from anhydride, C-O from alcohol and ether, and C-H stretching from alkanes [37]. For Cr^total adsorption by CASS-BA, gaps were found at 1035 and 1155 cm^-1, suggesting CO-O-CO stretching from anhydride and C-O stretching from secondary alcohol and aliphatic ether and C-H stretching from alkanes [61]; [37].

CASS-BB gaps are found for Cd²⁺ adsorption at 1733-1744 cm^-1 strong C=O stretching from conjugated anhydride, esters, and aldehyde and Pb²⁺ adsorption 1039 cm^-1, from medium C-H stretching from alkanes [38].

The pH_PZC values of the cassava natural adsorbents were $6.02\pm 0.06$ to CASS-BK, $6.04\pm 0.27$ to CASS-BA, and $6.26\pm 0.16$ to BASS-BB (Figure S1). Thus, for environment pH higher than the pH_PZC’s (6.02, 6.04, or 6.26), negative charges should predominate in the cassava adsorbent materials surface, favoring cation adsorption, such as heavy metal ions. Moreover, the CONAMA [39] guidance for natural water bodies states that natural waters tend to be found with pH values around 9 to 10, favoring heavy metals adsorption if these are in the aqueous medium.

Gonçalves Junior et al. [40], studying canola-based adsorbents, found values ranging from 5.95 to 6.04. Gonçalves Junior et al. [41], studying Euterpe Oleracea-based adsorbent, found 5.09 as pH_PZC. Schwantes et al. [42, 43] found 5.67 for a crambe-based adsorbent (Crambe abyssinica Hochst). Schwantes et al. [12] found 4.28 for the grape steam adsorbent. All the above authors conducted adsorption studies with pH values within the range 4.0 to 7.0, mainly due to the surface functional groups from lignin and cellulose structures that all the above adsorbents have in common and because heavy metals such as Cd²⁺, Pb²⁺, and Cr^total tend to precipitate in higher pH values.

When the $\text{p}{\text{H}}_{\text{PZC}}>\text{p}{\text{H}}_{\text{environment}}$ , the adsorbent surface is predominantly protonated (electropositive), favoring anion adsorption. On the other hand, when $\text{p}{\text{H}}_{\text{PZC}}<\text{p}{\text{H}}_{\text{environment}}$ the adsorbent surface is negatively charged, i.e., favoring cation adsorption [12].

The chemical composition of the cassava adsorbent (Table 1) is evidence around 22.5 to 35 g kg^-1 of Ca, 5.7 to 24 g kg^-1 of K, and 4.5 to 6.8 g kg^-1 of Mg. CASS-BK presents the higher concentration of Mn (123 mg kg^-1), while CASS-BA and CASS-BB present 27.6 and 34.0 mg kg^-1, respectively. Fe concentrations were found around 35.6 and 4.5 mg kg^-1, while Zn was found around 17.0 and 32.0 mg kg^-1.

Table 1

Composition of K, Ca, and Mg (g kg^-1) and Cu, Fe, Mn, Zn, Cd, Pb, and Cr (mg kg^-1) of the cassava adsorbents.

Adsorbent	K	Ca	Mg	Cu	Fe	Mn	Zn	Cd	Pb	Cr
	g kg-1					mg kg-1
CASS-BK	$24.10\pm 1.36$	$35.03\pm 1.36$	$6.83\pm 0.38$	$14.33\pm 0.57$	$35.67\pm 6.43$	$123.33\pm 1.52$	$32.00\pm 1.00$	<LQ	$11.00\pm 0.08$	<LQ
CASS-BA	$5.77\pm 0.49$	$23.23\pm 0.49$	$4.58\pm 0.10$	$5.67\pm 0.57$	$24.50\pm 4.50$	$27.67\pm 0.57$	$18.67\pm 0.57$	<LQ	$14.67\pm 0.57$	<LQ
CASS-BB	$7.77\pm 0.68$	$22.58\pm 0.69$	$5.12\pm 0.30$	$6.00\pm 1.00$	$26.00\pm 3.00$	$34.00\pm 13.89$	$17.00\pm 1.00$	<LQ	$3.33\pm 1.15$	<LQ

Notes: LQ: limits of quantification: $\text{K}=0.01$ , $\text{Ca}=0.005$ , $\text{Mg}=0.005$ , $\text{Cu}=0.005$ , $\text{Fe}=0.01$ , $\text{Mn}=0.01$ , $\text{Zn}=0.005$ , $\text{Cd}=0.005$ , $\text{Pb}=0.01$ , and $\text{Cr}=0.01$ (mg kg^-1).

Minor Pb levels were found in the cassava adsorbent composition, about 3.3 to 14.6 mg kg^-1. The lead is probably from (1) industrial contamination during the cassava roots processing, or (2) it was absorbed from the soil by the cassava plants and accumulated in the roots. Bassegio et al. [1] report that Pb is not translocated to upper vegetal organs, primarily accumulated in roots. This element was only detected because the adsorbent was submitted to strong acid digestion, which released the Pb bonded to the lignocellulosic structure of the adsorbent. Pb release in water samples was not observed.

3.2. Adsorption Studies (Cd²⁺, Pb²⁺, and Cr^total)

3.2.1. pH and Adsorbent Dose Studies

Table S3 shows the regression coefficients from Figure 3, while Table S4 shows the analysis of variance (ANOVA). Figure 4 illustrates the regression analysis for adsorbent doses. Finally, Table S3 shows the ideal pH and adsorbent dose conditions for the higher removal of heavy metals.

Figure 3

The influence of pH and the adsorbent doses on removing Cd²⁺, Pb²⁺, and Cr^total by CASS-BK, CASS-BA, and CASS-BB. Experimental conditions: adsorbent doses tested from 4.0 to 24 g L^-1, pH evaluated from 4.0 to 7.0, stirring at 160 rpm, at 298 K for 1.5 h.

Figure 4

Regression analysis for the adsorbent doses (4.0 to 24.0 g L^-1) and its performance in removing Cd²⁺, Pb²⁺, and Cr^total. Note: additional information regarding this figure is provided in Table S6.

The variable “Adsorbent” is significant at 1% of probability for Cd²⁺, Pb²⁺, and Cr^total removal, which means that there are differences in the removal rate of metal regarding the adsorbent (CASS-BK, CASS-BA, and CASS-BB). Furthermore, the parameter “Doses” is significant at 1% of probability for Cd²⁺, and 5% for Pb²⁺ removal, evidencing that the adsorbent doses influence heavy metal removal. Also, for Pb²⁺ removal, the variable “pH” of solution and the interaction between “Adsorbent” and “pH” presented statistical differences.

Tukey’s range test (Table S5) evidence the higher removal percentages at pH 5.0 for Cd²⁺ and Pb²⁺, while tested doses or pH values did not provoke statistical differences for Cr^total. Regarding the tested adsorbents, CASS-BB presented the higher removal of Cd²⁺ (98.8%). CASS-BA (98.6%) and CASS-BB (98.9%) had the highest Pb²⁺ removal percentages, while CASS-BB (93.7%) had the highest Cr^total removal percentage. The interaction between the evaluated factors “pH” and “dose” is provided in Table S7.

On the other hand, Tukey’s test range for the interaction between “Adsorbent” and “pH” (Table S7, Figure 3 and Figure S3) evidenced that CASS-BA presented higher means for Cd²⁺ removal at pH 6.0 (95.88%). CASS-BK presented higher means for Pb²⁺ removal at pH 4.0 and 5.0 and CASS-BA (99.96%) and CASS-BB (92.53%) at pH 5.0. Furthermore, CASS-BB presented higher removal of Cr^total at pH 4.0.

Figure S3(a) shows a comparison between the removal percentage of the heavy metals (mean of Cd²⁺, Pb²⁺, and Cr^total) and the cassava-based adsorbents, where the following order is observed: $\text{CASS}-\text{BB}\left(97.5\%\right)>\text{CASS}-\text{BA}\left(95.5\%\right)>\text{CASS}-\text{BK}\left(88.8\%\right)$ . Figure S3(b) shows the small variation observed for the averages of heavy metals removal in function to the pH evaluated range: $\text{pH}7.0\left(94.9\%\right)>\text{pH}5.0\left(94.3\%\right)>\text{pH}4.0\left(93.9\%\right)>\text{pH}6.0\left(92.9\%\right)$ . Figure S3(c) highlights the small variation for the averages of heavy metals removal in function to the adsorbent doses $:16\text{g}{\text{L}}^{-1}\left(94.3\%\right)>8.0\text{g}{\text{L}}^{-1}\left(94.2\%\right)=12\text{g}{\text{L}}^{-1}\left(94.2\%\right)>20\text{g}{\text{L}}^{-1}\left(94.1\%\right)=24\text{g}{\text{L}}^{-1}\left(94.1\%\right)>4.0\text{g}{\text{L}}^{-1}\left(92.7\%\right)$ .

Gonçalves Junior et al. [40] evaluated 16 canola-based adsorbents and reported the best adsorption rates using about 4.0 g L^-1 to remove Pb⁺². The above authors did not find significant variations in the studied pH range (4.0 to 7.0) on Pb²⁺ removal. However, Gonçalves Junior et al. [41] found a statistical difference for adsorbent doses and pH of Cd⁺² solution using Euterpe Oleracea endocarp-based adsorbent. The above authors report that 4.0 g L^-1 produced the higher removing percentage of Cd²⁺, Pb²⁺, and Cr^total from water, and, in the case of Cd²⁺ adsorption systems, the pH ranges from 6.0 to 7.0 presented higher removal rates. Similarly, Schwantes et al. [42, 43] found no pH dependency (from 3.0 to 7.0) on adsorption of Zn by crambe-based adsorbents, being the adsorbent dose the primary influencer on Zn removal. In their studies, the higher Zn removal rates were found using 5.0 g L^-1. On the other hand, Schwantes et al. [12] found statistical differences only for the adsorbent doses, with higher removal of Cd²⁺ found using 5.0 g L^-1 of grape steam-based adsorbent. These authors also claim that the evaluated pH range (3.0 to 7.0) did not influence Zn adsorption.

It is essential to state that heavy metal ions such as Cd, Pb, and Cr are only available for adsorption when dissolved in water solution, i.e., ions. Therefore, variations on the medium pH can cause the formation of complexes and favor their precipitation. For example, the theoretical diagram provided in Figure S2 illustrates the formation of different Cd, Pb, and Cr complexes in solution in $\text{pH}>7.0$ [22]. Thus, as this research is aimed at optimizing the conditions for adsorption of heavy metals and comparing cassava adsorbents to activated carbon, we only evaluated pH conditions that could not cause the complexation and precipitation of such substances.

3.2.2. Kinetics Studies

Figure 5 shows that the adsorption of Cd²⁺, Pb²⁺, and Cr^total by the cassava-based adsorbents is a relatively fast process, with 90% of the removal achieved within 5 minutes of contact adsorbent metal in solution, with the physicochemical equilibrium reached between 5 and 10 min. This result is an essential and technological advantage when the industrial use of cassava adsorbents is considered for wastewater treatment.

Figure 5

Kinetics of Cd²⁺ (a), Pb²⁺ (b), and Cr^total (c) adsorption by CASS-BK, CASS-BA, and CASS-BB. Experimental conditions: adsorbent dose of 8.0 g L^-1, pH of 5.5, stirring at 160 rpm, at 298 K, during 5, 10, 20, 40, 60, 80, 100, 120, 140, 160, and 180 min.

(b) (c)

The goodness of the fit obtained for PFO (Adj. $R^2$ , $R$ -square (COD), and reduced Chi²] suggests this model’s effectiveness in predicting the adsorption of heavy metals by the cassava-based adsorbents (Table 2). Nevertheless, there is a massive discrepancy by comparing the values of the amount of metal adsorbed estimated by the model ( $q_e$ calc.) and the experimentally obtained ( $q_e$ exp.).

Table 2

Kinetic parameters for the adsorption of Cd²⁺, Pb²⁺, and Cr^total by CASS-BK, CASS-BA, and CASS-BB, using the nonlinear models of pseudofirst order (PFO) and pseudosecond order (PSO).

Kinetic parameters	Cd2+			Pb2+			Crtotal
PFO	CASS-BK	CASS-BA	CASS-BB	CASS-BK	CASS-BA	CASS-BB	CASS-BK	CASS-BA	CASS-BB
$k_1$ (min-1)	$1.1928\pm 0.0026$	$1.2077\pm 0.0025$	$1.2425\pm 6.8616·{10}^{-4}$	$1.1453\pm 0.0063$	$1.2252\pm 0.0035$	$1.2308\pm 0.0038$	$1.0327\pm 0.0035$	$1.1601\pm 0.0050$	$1.1706\pm 0.0011$
$q_e$ cal. (mg g-1)	$0.5598\pm 0.0235$	$0.5656\pm 0.0229$	$0.8022\pm 0.0202$	$0.5313\pm 0.0517$	$0.5560\pm 0.0301$	$0.6789\pm 0.0619$	$0.6813\pm 0.0691$	$0.5870\pm 0.0536$	$0.8863\pm 0.0544$
Reduced Chi2	6.6307 10-5	6.08439 10-5	4.6745 10-6	0.0004	0.0002	0.0001	0.0001	0.0002	1.2741 10-5
$R$ -square (COD)	1.000	1.000	1.000	0.9967	1.000	0.9989	0.9987	1.000	0.9999
Adj. $R^2$	1.000	1.000	1.000	0.9964	1.000	0.9988	0.9986	1.000	0.9999
PSO	CASS-BK	CASS-BA	CASS-BB	CASS-BK	CASS-BA	CASS-BB	CASS-BK	CASS-BA	CASS-BB
$k_2$ (g mg-1 min-1)	$2.5019\pm 0.1252$	$2.5443\pm 0.1151$	$8.4765\pm 0.4883$	$1.8871\pm 0.1918$	$2.3549\pm 0.1940$	$3.8177\pm 0.6579$	$4.2999\pm 0.7287$	$2.4666\pm 0.2611$	$11.4556\pm 1.3146$
$q_e$ cal. (mg g-1)	$1.2012\pm 0.0012$	$1.2161\pm 0.0011$	$1.2451\pm 4.5838·{10}^{-4}$	$1.1577\pm 0.0032$	$1.2342\pm 0.0022$	$1.2372\pm 0.0029$	$1.0385\pm 0.0025$	$1.1698\pm 0.0027$	$1.1728\pm 6.8020·{10}^{-4}$
Reduced Chi2	1.1189 10-5	8.9005 10-6	1.5372 10-6	7.4262 10-5	3.3954 10-5	6.2170 10-5	4.6976 10-5	5.1007 10-5	3.3913 10-6
$R$ -square (COD)	0.9999	0.9999	0.9999	0.9994	0.9997	0.9995	0.9995	0.9996	0.9999
Adj. $R^2$	0.9999	0.9999	0.9999	0.9993	0.9997	0.9995	0.9995	0.9995	0.9999
$q_e$ exp. (mg g-1)	$1.1858\pm 0.022$	$1.2009\pm 0.022$	$1.2404\pm 0.007$	$1.1375\pm 0.030$	$1.2178\pm 0.025$	$1.2267\pm 0.017$	$1.0295\pm 0.015$	$1.1542\pm 0.024$	$1.1694\pm 0.005$

Notes: $q_e$ calc. and exp., respectively, the amount of metal adsorbed estimated by the model and obtained experimentally. $k_1$ : rate constant of pseudofirst order; $k_2$ : rate constant of pseudosecond order; $A$ : constant indicating the initial chemisorption rate; $B$ : number of suitable sites for adsorption, related to the extent of surface coverage and the activation energy of the chemo reaction; Reduced Chi²: residual sum of squares; $R$ -square (COD): coefficient of determination; Adj $R^2$ : adjusted coefficient of determination.

Therefore, PFO does not predict the observed phenomena correctly. On the other hand, the goodness of the fit obtained for PSO (Adj. $R^2$ , $R$ -square (COD), and reduced Chi²), and the excellent proximity between $q_e$ calc and $q_e$ exp strongly indicate that PSO efficiently predicts heavy metals’ adsorption by cassava-based adsorbents, being the most suitable model. Consequently, the removal of Cd²⁺, Pb²⁺, and Cr^total appear to be ruled by chemical forces, such as ionic or covalent bond processes [24].

Silveira Neta et al. [44] report that PFO and PSO models assume that the difference between the adsorbed concentration at a given time and the equilibrium adsorbed concentration is the driving force of adsorption. Furthermore, the global adsorption rate is either proportional to the driving force in the PFO equation or is the square of the driving force for the PSO model.

The PSO velocity constant ( $k_2$ expressed in g mg^-1 min^-1) presented the following values for Cd²⁺ adsorption: $k_{2\text{CASS}-\text{BB}}=8.47>k_{2\text{CASS}-\text{BA}}=2.54>k_{2\text{CASS}-\text{BK}}=2.50$ , while for Pb²⁺ adsorption the following order is observed: $k_{2\text{CASS}-\text{BB}}=3.82>k_{2\text{CASS}-\text{BA}}=2.35>k_{2\text{CASS}-\text{BK}}=1.88$ ; finally, for Cr^total adsorption, the following order is observed: $k_{2\text{CASS}-\text{BB}}=11.45>k_{2\text{CASS}-\text{BK}}=4.30>k_{2\text{CASS}-\text{BA}}=2.47$ . Regarding the estimated adsorbed amount ( $q_e$ calc., expressed in mg g^-1), for Cd²⁺ the following order is observed: $q_e\exp .{}_{\text{CASS}-\text{BB}}=1.24>q_e\exp .{}_{\text{CASS}-\text{BA}}=1.21>q_e\exp .{}_{\text{CASS}-\text{BK}}=1.20$ ; for Pb²⁺ adsorption: $q_e\exp .{}_{\text{CASS}-\text{BB}}=1.23=q_e\exp .{}_{\text{CASS}-\text{BA}}=1.23>q_e\exp .{}_{\text{CASS}-\text{BK}}=1.15$ ; and for Cr^total adsorption: $q_e\exp .{}_{\text{CASS}-\text{BB}}=1.17=q_e\exp .{}_{\text{CASS}-\text{BA}}=1.17>q_e\exp .{}_{\text{CASS}-\text{BK}}=1.04$ . Considering the above, CASS-BB is the fastest cassava adsorbent for heavy metal adsorption, with also the highest values of adsorbed amount.

Wu et al. [45] developed hydrothermal activated carbon from cassava slag biochar. The authors found $q_e$ of 24.10 mg g^-1 for Rhodamine-B removal, with $K_2\left(\text{g}\text{m}{\text{g}}^{-1}{\min }^{-1}\right)=0.019$ , i.e., much smaller than the lowest found in this study for Cd²⁺ ( $\text{CASS}-\text{BK}=2.50$ ), Pb²⁺ ( $\text{CASS}-\text{BK}=1.57$ ), and Cr^total ( $\text{CASS}-\text{BK}=1.10$ ). These results could evidence that cassava barks can be a more promising material in removing inorganic than organic contaminants. Moreover, our results evidence that CASS-BK is slower in adsorbing heavy metals. Chen et al. [46] found $K_2$ values (g mg^-1 min^-1) ranging from 0.0017 to 0.0059 using modified corn cobs and chestnut shells for Pb²⁺ removal. In both cases, adsorption velocity is much lower than observed in this research.

3.2.3. Equilibrium Studies and Side-by-Side Comparison (Activated Carbon) and the Reusability of the Materials in Other New Adsorption Cycles

Table S8 and S9 exhibit the ANOVA results, while the isotherms and $R^2$ values are presented in Table 3 and Figure S4. The goodness of the fit obtained for Freundlich, Langmuir, Temkin and Pyzhev, Khan, Sips, Liu, and Redlich-Peterson suggest good predictions of those models on the observed adsorption phenomena.

Table 3

Equilibrium parameters by the nonlinear models of Langmuir, Freundlich, Sips, Redlich-Peterson, Toth, Temkin, Liu, and Khan to remove Cd²⁺, Pb²⁺, and Cr^total by cassava-based adsorbents.

Parameters		CASS-BK	CASS-BA	CASS-BB	AC	CASS-BK	CASS-BA	CASS-BB	CASS-BK	CASS-BA	CASS-BB	AC
		Cd2+				Pb2+			Crtotal
Langmuir	$q_{\text{Langmuir}}$ (mg g-1)	$16.03158\pm 1.36538$	$14.74882\pm 0.86014$	$14.80667\pm 1.25349$	$21.49443\pm 0.53588$	$30.09417\pm 2.31339$	$25.25276\pm 1.24648$	$23.44436\pm 1,43539$	$18.3884\pm 2.01416$	$14.09721\pm 0.79578$	$13.28135\pm 0.38244$	$12.80244\pm 0.36463$
	$K_{\text{Langmuir}}$ (L mg-1)	$0.13825\pm 0.04738$	$0.11403\pm 0.02594$	$0.21865\pm 0.09085$	$0.57279\pm 0.04805$	$0.20149\pm 0.0351$	$0.52968\pm 0.07527$	$1.06964\pm 0.21915$	$0.04017\pm 0.0106$	$0.04855\pm 0.00811$	$0.08824\pm 0.00918$	$0.84959\pm 0.14625$
	$R_{\text{Langmuir}}$ (5 to 200 mg L-1)	0.035 to 0.591	0.042 to 0.637	0.022 to 0.478	0.009 to 0.259	0.024 to 0.498	0.009 to 0.274	0.005 to 0.158	0.111 to 0.833	0.093 to 0.805	0.054 to 0.694	0.006 to 0.191
	Reduced Chi2	2.04954	0.82442	2.65031	0.3994	1.11639	1.10101	2.32819	1.28388	0.33446	0.15877	0.51943
	$R$ -square (COD)	0.93203	0.96577	0.90632	0.99261	0.98171	0.98234	0.9635	0.94969	0.98036	0.99129	0.97845
	Adj. $R^2$	0.9235	0.96149	0.89461	0.99169	0.97942	0.98013	0.95894	0.9434	0.97791	0.9902	0.97575
Freundlich	$K_{\text{freundlich}}$ (mg g−1 (mg L−1)−1/n )	$3.63599\pm 0.28729$	$3.30651\pm 0.24546$	$4.96969\pm 0.18872$	$7.95208\pm 0.91248$	$5.78945\pm 0.51863$	$8.40103\pm 0.70876$	$10.32826\pm 0.73736$	$1.90101\pm 0.6364$	$1.94262\pm 0.38241$	$2.68319\pm 0.51933$	$5.69459\pm 0.62652$
	$n_{\text{Freundlich}}$ (dimensionless)	$2.80628\pm 0.17702$	$2.93935\pm 0.17473$	$3.76663\pm 0.15428$	$3.17685\pm 0.46251$	$1.84867\pm 0.14726$	$2.35796\pm 0.24608$	$2.65845\pm 0.29211$	$2.15719\pm 0.41389$	$2.47454\pm 0.30651$	$2.90776\pm 0.42751$	$4.70953\pm 0.6928$
	$1/n_{\text{Freundlic}}$ h (dimensionless)	0.3563	0.3402	0.2655	0.315	0.5409	0.4241	0.3762	0.4636	0.4041	0.3439	0.2123
	Reduced Chi2	0.42388	0.30506	0.1459	5.8525	1.71008	3.34026	3.71903	3.81642	0.96709	1.54048	2.06817
	$R$ -square (COD)	0.98594	0.98733	0.99484	0.89173	0.97198	0.94641	0.9417	0.85045	0.94322	0.9155	0.91419
	Adj. R 2	0.98418	0.98575	0.9942	0.8782	0.96848	0.93971	0.93441	0.83175	0.93612	0.90493	0.90346
Sips	$q_{\text{Sips}}$ (mg g-1)	$65.38617\pm 81.16848$	$25.39482\pm 4.38679$	$131.82648\pm 247.93561$	$21.35258\pm 0.87117$	$37.41387\pm 10.90253$	$27.79766\pm 5.07839$	$26.99272\pm 3.36206$	$14.7634\pm 1.23577$	$14.79742\pm 2.13484$	$13.20523\pm 0.74521$	$14.00721\pm 0.55321$
	$K_{\text{Sips}}$ (L mg-1)	$9.8368710-4\pm 0.00441$	$0.01926\pm 0.01263$	$1.1398410-5\pm 9.638010-5$	$0.58341\pm 0.07056$	$0.12015\pm 0.08481$	$0.63578\pm 0.37545$	$0.44236\pm 0.15258$	$0.05997\pm 0.009$	$0.04348\pm 0.01608$	$0.08935\pm 0.01365$	$0.62075\pm 0.11802$
	$n_{\text{Sips}}$	$0.41406\pm 0.09095$	$0.52448\pm 0.05307$	$0.28496\pm 0.04239$	$1.01783\pm 0.09192$	$0.84556\pm 0.15122$	$0.76903\pm 0.16365$	$0.90902\pm 0.14204$	$1.59095\pm 0.28295$	$0.92013\pm 0.1756$	$1.01434\pm 0.11589$	$0.69058\pm 0.0708$
	Reduced Chi2	0.45231	0.10491	0.16054	0.45392	1.1103	2.16475	1.19517	0.73073	0.37173	0.18103	0.20415
	$R$ -square (COD)	0.98687	0.99619	0.99503	0.99265	0.98408	0.97031	0.98322	0.97494	0.9809	0.99131	0.99259
	Adj. $R^2$	0.98312	0.9951	0.99362	0.99055	0.97953	0.96182	0.97843	0.96779	0.97545	0.98883	0.99047
Redlich-Peterson	$K_{\text{RP}}$ (L g-1)	$16.70571\pm 12.37434$	$8.36071\pm 2.54453$	$283.25074\pm 224.62529$	$13.26416\pm 1.67146$	$8.32751\pm 2.94818$	$38.72759\pm 14.2866$	$16.7139\pm 4.08174$	$0.43367\pm 0.02326$	$0.59042\pm 0.14514$	$1.01102\pm 0.14273$	$18.64749\pm 3.19075$
	$a_{\text{RP}}$ (L mg-1)	$3.72107\pm 3.25994$	$1.84552\pm 0.71026$	$54.4878\pm 44.66023$	$0.67642\pm 0.16112$	$0.52827\pm 0.43438$	$2.3358\pm 1.27182$	$0.89245\pm 0.41738$	$2.29001{10}^{-4}\pm 2.1952·{10}^{-4}$	$0.02661\pm 0.02844$	$0.05474\pm 0.02478$	$2.01019\pm 0.45332$
	${ᵦ}_{\text{RP}}$ (dimensionless)	$0.69553\pm 0.03834$	$0.73312\pm 0.02143$	$0.74687\pm 0.01166$	$0.96965\pm 0.03829$	$0.77516\pm 0.15977$	$0.83509\pm 0.0845$	$0.87888\pm 0.09025$	$2.01047\pm 0.19961$	$1.09409\pm 0.17562$	$1.07179\pm 0.07031$	$0.91411\pm 0.01727$
	Reduced Chi2	0.32188	0.07992	0.10797	0.41509	0.97008	1.72126	0.97036	0.17667	0.37164	0.15871	0.14841
	$R$ -square (COD)	0.99066	0.9971	0.99666	0.99328	0.98609	0.97639	0.98638	0.99394	0.98091	0.99238	0.99461
	Adj. R2	0.98799	0.99627	0.99571	0.99136	0.98212	0.96964	0.98249	0.99221	0.97545	0.99021	0.99307
Toth	$q_{\text{Toth}}$ (mg g-1)	$392.17758\pm 1418.39671$	$46.32464\pm 19.32673$	$61269.30629\pm 1.07973E6$	$21.76049\pm 1.15496$	$47.44475\pm 28.74734$	$31.47671\pm 9.72964$	$29.51154\pm 6.0196$	$13.34476\pm 0.52064$	$14.23303\pm 2.60567$	$12.92332\pm 0.85817$	$14.62388\pm 0.75137$
	$b_t$ (dimensionless)	$17.79585\pm 93.2196$	$1.14213\pm 0.86741$	$1.83899E11\pm 1.07978E13$	$0.59026\pm 0.09258$	$0.19285\pm 0.04945$	$1.57452\pm 0.77771$	$0.59327\pm 0.14267$	$0.02969\pm 0.00183$	$0.04924\pm 0.01337$	$0.08228\pm 0.01424$	$2.29449\pm 0.81443$
	$n_{\text{Toth}}$ (dimensionless)	$0.11468\pm 0.10939$	$0.24313\pm 0.06156$	$0.0355\pm 0.07232$	$0.95864\pm 0.1546$	$0.60065\pm 0.30642$	$0.56938\pm 0.24045$	$0.74607\pm 0.23928$	$5.35184\pm 2.51761$	$0.97637\pm 0.38958$	$1.09872\pm 0.23239$	$0.54805\pm 0.07842$
	Reduced Chi2	0.41282	0.08771	0.05805	0.45161	1.02201	1.96881	1.0885	0.33913	0.38211	0.17679	0.16318
	$R$ -square (COD)	0.98802	0.99681	0.99994	0.99269	0.98535	0.97299	0.98472	0.98837	0.98037	0.99151	0.99408
	Adj. $R^2$	0.9846	0.9959	0.99992	0.9906	0.98116	0.96528	0.98035	0.98505	0.97476	0.98909	0.99238
Temkin and Pyzhev	$b_{\text{TP}}$ (dimensionless)	$1.00569\pm 0.07297$	$1.14545\pm 0.08256$	$1.5761\pm 0.15064$	$0.61398\pm 0.02149$	$0.44159\pm 0.02692$	$0.54628\pm 0.02374$	$0.4846\pm 0.00996$	$0.63938\pm 0.05938$	$1.05262\pm 0.0929$	$1.02271\pm 0.06056$	$1.37757\pm 0.05278$
	$A_{\text{TP}}$ (L mg-1)	$5.19005\pm 1.53976$	$5.19616\pm 1.6121$	$74.10073\pm 44.81592$	$7.75829\pm 0.84755$	$3.06553\pm 0.43124$	$14.36746\pm 1.82544$	$6.45312\pm 0.34133$	$0.44175\pm 0.09563$	$1.17421\pm 0.34885$	$1.5298\pm 0.3132$	$28.33647\pm 5.84206$
	Reduced Chi2	1.20705	0.95169	1.90872	0.51953	1.74482	0.94036	0.20794	1.63193	0.98971	0.49258	0.27695
	$R$ -square (COD)	0.95997	0.96048	0.93253	0.99039	0.97141	0.98526	0.99666	0.93605	0.94189	0.97298	0.98851
	Adj. $R^2$	0.95496	0.95554	0.9241	0.98919	0.96784	0.98342	0.99625	0.92806	0.93463	0.9696	0.98707
Liu	$q_{\text{Liu}}$ (mg g-1)	$65.38617\pm 81.16848$	$25.39482\pm 4.38679$	$131.82648\pm 247.93561$	$21.35258\pm 0.87117$	$37.41387\pm 10.90253$	$27.79766\pm 5.07839$	$26.99272\pm 3.36206$	$14.7634\pm 1.23577$	$14.79742\pm 2.13484$	$13.20523\pm 0.74521$	$14.00721\pm 0.55321$
	$k_g$ (L mg-1)	$9.8368710-4\pm 0.00441$	$0.01926\pm 0.01263$	$1.13984{10}^{-5}\pm 9.6380{10}^{-5}$	$0.58341\pm 0.07056$	$0.12015\pm 0.08481$	$0.63578\pm 0.37545$	$0.44236\pm 0.15258$	$0.05997\pm 0.009$	$0.04348\pm 0.01608$	$0.08935\pm 0.01365$	$0.62075\pm 0.11802$
	$n_{\text{Liu}}$ (dimensionless)	$0.41406\pm 0.09095$	$0.52448\pm 0.05307$	$0.28496\pm 0.04239$	$1.01783\pm 0.09192$	$0.84556\pm 0.15122$	$0.76903\pm 0.16365$	$0.90902\pm 0.14204$	$1.59095\pm 0.28295$	$0.92013\pm 0.1756$	$1.01434\pm 0.11589$	$0.69058\pm 0.0708$
	Reduced Chi2	0.45231	0.10491	0.16054	0.45392	1.1103	2.16475	1.19517	0.73073	0.37173	0.18103	0.20415
	$R$ -square (COD)	0.98687	0.99619	0.99503	0.99265	0.98408	0.97031	0.98322	0.97494	0.9809	0.99131	0.99259
	Adj. $R^2$	0.98312	0.9951	0.99362	0.99055	0.97953	0.96182	0.97843	0.96779	0.97545	0.98883	0.99047
Khan	$q_{\text{Khan}}$ (mg g-1)	$2.79776\pm 1.09417$	$3.39747\pm 0.65966$	$1.58272\pm 0.4133$	$18.78451\pm 2.71831$	$11.86292\pm 7.7197$	$11.6408\pm 4.5597$	$15.33791\pm 5.38731$	$308383.429\pm 6.50488{10}^8$	$23.82914\pm 14.28871$	$17.93275\pm 3.93123$	$8.38776\pm 0.85235$
	$b_{\text{Khan}}$ (L mg-1)	$3.66843\pm 3.1002$	$1.86396\pm 0.79712$	$103.28289\pm 86.05213$	$0.70103\pm 0.16585$	$0.64488\pm 0.5585$	$3.01044\pm 1.93491$	$1.04077\pm 0.53624$	$1.75541{10}^{-6}\pm 0.0037$	$0.02429\pm 0.01851$	$0.05697\pm 0.01811$	$2.0356\pm 0.49317$
	$a_{\text{Khan}}$ (dimensionless)	$0.68337\pm 0.0296$	$0.71157\pm 0.01807$	$0.74516\pm 0.01046$	$0.95377\pm 0.04581$	$0.66711\pm 0.14614$	$0.79328\pm 0.08209$	$0.82211\pm 0.09717$	$8129.76185\pm 1.7233107$	$1.31273\pm 0.46136$	$1.14086\pm 0.11927$	$0.90671\pm 0.01841$
	Reduced Chi2	0.27699	0.08228	0.0955	0.39659	0.91343	1.60777	0.89426	0.25455	0.35402	0.14764	0.15629
	$R$ -square (COD)	0.99196	0.99701	0.99705	0.99358	0.9869	0.97795	0.98745	0.99127	0.98181	0.99291	0.99433
	Adj. $R^2$	0.98967	0.99616	0.9962	0.99175	0.98316	0.97165	0.98386	0.98878	0.97662	0.99089	0.9927

Notes: it was impossible to construct an isotherm for the comparison study of Pb²⁺ with activated carbon (AC).

(1) Cd²⁺ Equilibrium Studies. Langmuir predictions indicate a higher theoretical adsorption capacity ( $q_{\text{Langmuir}}$ , expressed in mg g^-1) for the AC (with $q_{\text{Langmuir}}=21.49$ ), followed by CASS-Bk ( $q_{\text{Langmuir}}=16.03$ ), CASS-BB ( $q_{\text{Langmuir}}=14.81$ ), and CASS-BA ( $q_{\text{Langmuir}}=14.75$ ). Furthermore, $K_{\text{Langmuir}}$ values (expressed in L mg^-1) also indicate higher affinity between AC and Cd²⁺ than cassava-based adsorbents, with the following order: $K_{\text{Langmuir}\text{AC}}0.57>K_{\text{Langmuir}\text{CASS}-\text{BB}}\text{of}0.22>K_{\text{Langmuir}\text{CASS}-\text{BK}}\text{of}0.13>K_{\text{Langmuir}\text{CASS}-\text{BA}}\text{of}0.11$ . Thus, AC presented higher maximum adsorption capacity and affinity with Cd²⁺ than cassava adsorbents.

Similar behavior is seen in Freundlich parameters, with $K_{\text{Frelundlich}}$ (expressed in (mg g⁻¹ (mg L⁻¹)^−1/n )) presenting the following order: $K_{\text{Freundlich}\text{AC}}\text{of}7.95>K_{\text{Freundlich}\text{CASS}-\text{BB}}\text{of}4.96>K_{\text{Freundlich}\text{CASS}-\text{BK}}\text{of}3.63>K_{\text{Freundlich}\text{CASS}-\text{BA}}\text{of}3.31$ . Nevertheless, $n_{\text{Freundlich}\text{CASS}-\text{BB}}\left(3.76\right)>n_{\text{Freundlich}\text{AC}}\left(3.18\right)$ indicates a higher adsorption intensity and a better distribution of the energy and heterogeneity of adsorbate sites [47] for CASS-BB against AC. Moreover, $n_{\text{Freundlich}}$ values are found between 1 and 10, and $1/n$ is between 0 and 1, indicating a favorable and cooperatively adsorption [48–50].

Sips, also known as the Langmuir-Freundlich model, can predict simultaneously mono and multilayer adsorption. For example, CASS-BK (65.38) and CASS-BB (131.82) presented higher $q_{\text{Sips}}$ (expressed in mg g^-1) than AC (21.35), i.e., predicting that cassava-based adsorbents could remove more Cd²⁺ from water than AC in the evaluated physicochemical conditions. Nevertheless, $K_{\text{Sips}}$ values show that AC possesses a higher affinity with Cd²⁺ than cassava-adsorbents ( $K_{\text{Sips}\text{AC}}>K_{\text{Sips}\text{CASSAVA}-\text{ADSORBENTS}}$ ). Furthermore, the cassava adsorbents presented $n_{\text{Sips}}$ lower than 0.6 (i.e., deviation from the unity), indicating higher proximity to the Freundlich model, while AC presented $n_{\text{Sips}}=1.018$ , suggesting higher similarity to the Langmuir assumptions [51].

The Redlich-Peterson model is similar to Sips, as it is a hybrid isotherm model that features Freundlich and Langmuir isotherm models with three parameters [52]. It approaches the Freundlich model at high concentration (as ${\beta }_{RP}$ →0) and at low concentration to Langmuir isotherm (as ${\beta }_{RP}$ →1) [53]. The values of ${\beta }_{RP}$ follow the order: ${\beta }_{RP-\text{AC}}$ (0.969) > ${\beta }_{RP}$ _CASS-BB (0.744) > ${\beta }_{RP}$ _CASS-BA (0.733) > ${\beta }_{RP}$ _CASS-BK (0.695). As evidenced above, $n_{\text{Sips}}$ assumes values inferior but closer to the unity (exception for AC, with ${\beta }_{RP}>1$ ), indicating certain proximity with the Langmuir monolayer model.

The b_TP values (dimensionless) obtained by the Temkin model follow the order: $b_{\text{TP}\text{BB}}\left(1.57\right)>b_{\text{TP}\text{CASS}-\text{BA}}\left(1.14\right)>b_{\text{TP}\text{CASS}-\text{BK}}\left(1.00\right)>b_{\text{TP}\text{AC}}\left(0.61\right)$ , which could be assigned to the more vital interaction between Cd²⁺ ions and active sites of the cassava adsorbents when compared to AC [54].

Liu’s model is derived from a combination of Langmuir and Freundlich models; thus, the Langmuir assumption of the monolayer formation and the Freundlich infinite adsorption assumption are eliminated by the Liu model [55]. This model predicts that active adsorbent sites cannot have the same energy, and its often used for better predictions of Liu’s maximum adsorption capacity prediction ( $q_{\text{Liu}}$ ). The $q_{\text{Liu}}$ (mg g^-1) predicts the following order: $q_{\text{Liu}\text{CASS}-\text{BB}}\left(131.82\right)>q_{\text{Liu}\text{CASS}-\text{BK}}\left(65.38\right)>q_{\text{Liu}\text{CASS}-\text{BA}}\left(25.39\right)>q_{\text{Liu}\text{CASS}-\text{AC}}\left(21.35\right)$ , suggesting higher removal of Cd²⁺ by the cassava-adsorbents than AC.

Khan’s predictions on Cd²⁺ maximum adsorption capacity ( $q_{\text{Khan}}$ , expressed in mg g^-1) followed the order: $q_{\text{Khan}\text{AC}}\left(18.78\right)>q_{\text{Khan}\text{BB}}\left(3.98\right)>q_{\text{Khan}\text{BA}}\left(3.39\right)>q_{\text{Khan}\text{BK}}\left(2.79\right)$ , indicating superiority of AC in Cd²⁺ removal. When the $a_{\text{Khan}}$ exponent equals 1, Khan isotherm approaches Langmuir monolayer interpretation [56], as observed for $a_{\text{Khan}}\text{AC}=0.953$ , but not for cassava adsorbents, with $0.68<a_{\text{Khan}\text{adsorbents}}<0.74$ .

(2) Pb²⁺ Equilibrium Studies. CASS-BK presented the highest Pb²⁺ adsorption maximum capacity ( $q_{\text{Langmuir}}=30.09$  mg g^-1), followed by CASS-BA ( $q_{\text{Langmuir}}=25.25$  mg g^-1), and CASS-BB ( $q_{\text{Langmuir}}=23.44$  mg g^-1). Nevertheless, CASS-BB presented the highest K_Langmuir values despite lower maximum adsorption capacity, indicating a higher affinity between Pb²⁺ and CASS-BB against other cassava adsorbents.

$K_{\text{Freundlich}}$ values ((mg g⁻¹ (mg L⁻¹)^−1/n )) indicate a higher affinity of CASS-BB with Pb²⁺ ( $K_{\text{Freundlich}\text{CASS}-\text{BB}}=10.32$ ), than CASS-BA ( $K_{\text{Freundlich}\text{CASS}-\text{BA}}=8.40$ ), and CASS-BK ( $K_{\text{Freundlich}\text{CASS}-\text{BK}}=5.79$ ). Moreover, $n_{\text{Freundlich}}$ values suggest the same order for Pb²⁺ adsorption intensity $\text{CASS}-\text{BB}>\text{CASS}-\text{BA}>\text{CASS}-\text{BK}$ .

The maximum capacity of Pb²⁺ adsorption predicted by Sips (expressed in mg g^-1) followed the order: $q_{\text{Sips}\text{CASS}-\text{BK}}\text{of}37.41>q_{\text{Sips}\text{CASS}-\text{BB}}\text{of}27.79>q_{\text{Sips}\text{CASS}-\text{BA}}\text{of}26.99$ . $n_{\text{Sips}}$ values are lower than 1 ( $0.12<q_{\text{Sips}}<0.63$ ), suggesting proximity with Freundlich multilayer assumptions. $K_{\text{Sips}}$ (L mg^-1) values indicate the following order of affinity with Pb²⁺: $K_{\text{Sips}\text{CASS}-\text{BA}}\text{of}0.63>K_{\text{Sips}\text{CASS}-\text{BB}}\text{of}0.44>K_{\text{Sips}\text{BK}}\text{of}0.12$ .

Using activated carbon (AC) from tobacco wastes for the removal of Pb²⁺, [40] found $q_{\text{Langmuir}}=71.573$  mg g^-1, $K_{\text{Freundlich}}=76.076$ ((mg g⁻¹ (mg L⁻¹)^−1/n )), and $q_{\text{Sips}}=155.440$ (mg g^-1), i.e., higher than the observed for cassava-based adsorbents. Nevertheless, it is important to consider the enhanced characteristics of an AC compared to an unmodified adsorbent. Furthermore, although the lower adsorption performance, cassava adsorbents have the advantage of easy obtaining and low cost of production.

K_RP (L g^-1) values followed the order: $K_{\text{RP}\text{CASS}-\text{BA}}\text{of}38.72>K_{\text{RP}\text{CASS}-\text{BB}}\text{of}16.71>K_{\text{RP}\text{CASS}-\text{BK}}\text{of}8.32$ , while ${\beta }_{RP}$ values follow the order $:{\beta }_{RP\text{BB}}\text{of}0.878>{\beta }_{RP\text{CASS}-\text{BA}}\text{of}0.835>{\beta }_{RP\text{CASS}-\text{BK}}\text{of}0.775$ , i.e., inferior but closer to 1, indicating proximity to the Langmuir model. Temkin $b_{\text{TP}}$ values follow the order: $b_{\text{TP}\text{CASS}-\text{BA}}\text{of}0.54>b_{\text{TP}\text{CASS}-\text{BB}}\text{of}0.48>b_{\text{TP}\text{CASS}-\text{BK}}\text{of}0.44$ , suggesting a higher interaction between Pb²⁺ and CASS-BA than CASS-BK [54]. The $q_{\text{Liu}}$ (mg g^-1) obtained for Pb²⁺ removal predicts the following order: $q_{\text{Liu}\text{CASS}-\text{BK}}\text{of}37.38>q_{\text{Liu}\text{CASS}-\text{BA}}\text{of}27.78>q_{\text{Liu}\text{CASS}-\text{BB}}\text{of}26.98$ , suggesting higher removal of Pb²⁺ by CASS-BK than CASS-BB [55].

The following order is observed for $q_{\text{Khan}}$ parameter ( $q_{\text{Khan}}$ , expressed in mg g^-1): $q_{\text{Khan}\text{CASS}-\text{BB}}\text{of}15.33>q_{\text{Khan}\text{CASS}-\text{BK}}\text{of}11.86>q_{\text{Khan}\text{CASS}-\text{BA}}\text{of}11.6$ 4, indicating a higher adsorption capacity prediction for CASS-BB. Furthermore, similar to Cd²⁺ adsorption, a_Khan exponent was distant from the unity, suggesting more excellent proximity to Freundlich assumptions [56].

(3) Cr^total Equilibrium Studies. Langmuir model predicted the following values for maximum Cr^total adsorption (q_Langmuir, expressed in mg g^-1): q_{Langmuir CASS-BK} (18.38) > q_{Langmuir CASS-BA} (14.09) > q_{Langmuir CASS-BB} (13.28) > q_{Langmuir AC} (12.80), illustrating a superiority of cassava adsorbents compared to AC. Furthermore, $K_{\text{Langmuir}}$ values demonstrate that CASS-BB is as good as AC regarding chemical affinity with Cr^total, with the higher $K_{\text{Langmuir}\text{CASS}-\text{BB}}\left(0.882\right)>K_{\text{Langmuir}\text{CASS}-\text{AC}}\left(0.849\right)>K_{\text{Langmuir}\text{CASS}-\text{BA}}\left(0.048\right)>K_{\text{Langmuir}\text{CASS}-\text{BK}}\left(0.040\right)$ .

Nevertheless, $K_{\text{Freundlich}}$ ((mg g⁻¹ (mg L⁻¹)^−1/n )) predict a higher affinity of AC–Cr^total adsorption: $K_{\text{Freundlich}\text{AC}}\left(5.69\right)>K_{\text{Freundlich}\text{CASS}-\text{BB}}\left(2.68\right)>K_{\text{Freundlich}\text{CASS}-\text{BA}}\left(1.94\right)>K_{\text{Freundlich}\text{CASS}-\text{BK}}\left(1.90\right)$ . Similarly, in Cd²⁺ and Pb²⁺ adsorption studies, $n_{\text{Freundlich}}$ presented values within 1 and 10, and $1/n_{\text{Freundlich}}$ between 0 and 1 ( $2.15<n_{\text{Freundlich}}<4.71;0.21<1/n_{\text{Freundlich}}<0.46$ ), suggesting a favorable and cooperatively adsorption system [48–50].

$q_{\text{Sips}}$ (mg g^-1) shows a consistent behavior, with certain proximity between the Cr^total maximum adsorption capacities, and a slight advantage for CASS-BA against AC: $q_{\text{Sips}\text{CASS}-\text{BA}}\left(14.79\right)>q_{\text{Sips}\text{CASS}-\text{BK}}\left(14.76\right)>q_{\text{Sips}\text{CASS}-\text{AC}}\left(14.00\right)>q_{\text{Sips}\text{CASS}-\text{BB}}\left(13.20\right)$ . Furthermore, $n_{\text{Sips}}<1$ suggests proximity to Freundlich assumptions for CASS-BK ( $n_{\text{Sips}}=0.91$ ) and AC ( $n_{\text{Sips}}=0.69$ ), while a proximity to Langmuir is observed for CASS-BK ( $n_{\text{Sips}}=1.87$ ) and CASS-BB ( $n_{\text{Sips}}=1.01$ ).

The Redlich-Peterson exponent ( ${\beta }_{RP}$ ) presented the following values: ${\beta }_{RP}\text{CASS}-\text{BK}=2.01>{\beta }_{RP}\text{CASS}-\text{BA}=1.09>{\beta }_{RP}\text{CASS}-\text{BA}=1.07>{\beta }_{RP}\text{CASS}-\text{AC}=0.91$ , i.e., cassava adsorbents are mostly ${\beta }_{RP}$ →1, suggesting proximity to the Langmuir model, while the opposite is observed for AC.

For Cr^total adsorption, b_TP values follow the order: $b_{\text{TP}\text{CASS}-\text{AC}}\left(1.37\right)>b_{\text{TP}\text{CASS}-\text{BA}}\left(1.05\right)>b_{\text{TP}\text{CASS}-\text{BB}}\left(1.022\right)>b_{\text{TP}\text{CASS}-\text{BK}}\left(0.64\right)$ . Thus, indicating higher interaction between Cr^total and AC than the cassava adsorbents. It is essential to highlight that CASS-BK presented the higher values of $b_{\text{TP}}$ for Pb²⁺ but had the lowest values for Cr^total adsorption.

The maximum adsorption of Cr^total predicted by the Liu model indicates the following order: $q_{\text{Liu}\text{CASS}-\text{BA}}\left(14.80\right)>q_{\text{Liu}\text{CASS}-\text{BK}}\left(14.76\right)>q_{\text{Liu}\text{CASS}-\text{AC}}\left(14.00\right)>q_{\text{Liu}\text{CASS}-\text{BB}}\left(13.20\right)$ , highlighting a superiority of CASS-BA against AC.

$q_{\text{Khan}}$ shows maximum adsorption capacity predictions for Cr^total in the following order: $q_{\text{Khan}\text{CASS}-\text{BA}}\left(23.82\right)>q_{\text{Khan}\text{BB}}\left(17.93\right)>q_{\text{Khan}\text{CASS}-\text{AC}}\left(8.39\right)$ . It is essential to state that the results of CASS-BK were overestimated. Nevertheless, the results suggest the superiority of CASS-BA and CASS-BB against AC in Cr^total adsorption. Furthermore, a_Khan exponent was superior to 1 for CASS-BA and CASS-BB, suggesting proximity to Langmuir assumptions, while for AC $a_{\text{Khan}}<1$ , suggesting higher Freundlich proximity [56].

3.2.4. The Effect of Temperature on Cd²⁺, Pb²⁺, and Cr^total adsorption by the Cassava Adsorbents

The effect of temperature on heavy metal adsorption is shown in Figure 6, with $q_e$ values increasing with temperature rise. On the other hand, entropy, enthalpy, and Gibbs’s free energy decrease with temperature increase.

Figure 6

Thermodynamic parameters obtained in the adsorption of Cd²⁺ (a), Pb²⁺ (b), and Cr^total (c) by CASS-BK, CASS-BA, and CASS-BB. Experimental conditions: adsorbent dose of 8.0 g L^-1, pH of 5.5, stirring at 160 rpm for 60 min, $C_0$ of Cd²⁺, Pb²⁺, and $\text{C}{\text{r}}^{\text{total}}=50$ , 100, and 50 mg L^-1.

(b) (c)

For Cd²⁺ adsorption, $∆H^{°}$ (kJ mol^-1) values followed the order $:\text{CASS}.\text{BA}\left(28.56\right)>\text{CASS}.\text{BB}\left(27.47\right)>\text{CASS}.\text{BK}\left(5.56\right);\text{for}\text{P}{\text{b}}^{2+}\text{adsorption}:\text{CASS}.\text{BB}\left(33.81\right)>\text{CASS}.\text{BA}\left(27.20\right)>\text{CASS}.\text{BK}\left(-10.01\right)$ , while for Cr^total adsorption: $\text{CASS}.\text{BB}\left(2.63\right)>\text{CASS}.\text{BA}\left(20.94\right)>\text{CASS}.\text{BK}\left(23.69\right)$ . Therefore, only CASS.BK presented exothermic adsorption of Pb²⁺, while the other studied cases the heavy metal adsorption is endothermic.

For Cd²⁺ adsorption, $∆S^{°}$ (J mol^-1) values assumed the following order: $\text{CASS}.\text{BB}\left(95.13\right)>\text{CASS}.\text{BA}\left(89.62\right)>\text{CASS}.\text{BK}\left(20.03\right)$ , for Pb²⁺ adsorption: $\text{CASS}.\text{BB}\left(125.94\right)>\text{CASS}.\text{BA}\left(99.98\right)>\text{CASS}\text{BK}\left(35.00\right)$ , while for Cr^total adsorption: $\text{CASS}.\text{BB}\left(75.06\right)>\text{CASS}.\text{BA}\left(61.59\right)>\text{CASS}.\text{BK}\left(2.08\right)$ . Therefore, higher randomness is observed using the bagasse of cassava (BA) or its mixture with barks (BB) when compared to cassava barks (BK).

3.2.5. Heavy Metal Removal by Cassava-Based Adsorbents and Desorption Studies—General Considerations, the Summary of Our Findings

Figure 7(d) shows 83% of adsorption of heavy metals (Cd²⁺, Pb²⁺, and Cr^total) by the cassava-based materials, with a 47% rate of desorption. On the other hand, the AC presented 93% adsorption with a 53% desorption rate. In other words, AC presented higher adsorption and desorption of heavy metals. Nevertheless, cassava-based adsorbents presented close results, highlighting their great potential.

Figure 7

Cd²⁺ (a), Pb²⁺ (b), and Cr^total (c) adsorption (ADS) and desorption (DES) percentage for CASS-BK, CASS-BA, and CASS-BB, in different initial concentrations (from 5.0 to 200 mg L^-1). Mean for the adsorption and desorption of 3 heavy metals (d). Experimental conditions: adsorbent dose of 8.0 g L^-1, pH of 5.5, stirring at 160 rpm for 60 min, at 298 K.

(b) (c) (d)

Wu et al. [45] removed 96% of Rhodamine B using cassava slag biochar. Guo et al. [57] removed 99% of Cr (IV) using cassava sludge-based AC. Chen et al. [46] reached 79% and 86% of Pb²⁺ removal using modified corn cobs and chestnut shells. Scheufele et al. [58] removed 60% of direct black dye using cassava root husks. As observed, cassava-based adsorbents and their derivatives still have great potential for removing organic and inorganic contaminants from water mediums.

As mentioned above, the adsorption means obtained during the equilibrium studies highlight the great potential of cassava-adsorbents removing heavy metals, such as Cd²⁺, Pb²⁺, and Cr^total from water. CASS-BK removed on average 83.98% of Cd²⁺, 96.06% of Pb²⁺, and 65.65% of Cr^total; CASS-BA removed 80.27% of Cd²⁺, 97.36% of Pb²⁺, and 67.76% of Cr^total; CASS-BB removed 84.91% of Cd²⁺, 98.18% of Pb²⁺, and 73.72% of Cr^total (Figure 7); meanwhile, the AC, a commercial adsorbent removed 95.74% of Cd²⁺, 99.43% of Pb²⁺, and 84.65% of Cr^total. Thus, considering the averages obtained for the removal of heavy metal combined, cassava-adsorbents presented $83.10\%\pm 12.44$ of removal, against $93.27\%\pm 7.69$ from AC. Therefore, cassava-based adsorbents are competitive alternatives, especially considering their high availability and low cost.

The acid elution with HCl 0.1 M recovered a significant part of Cd²⁺ and Pb²⁺, but only a fraction of Cr^total adsorbed by cassava materials. The desorption procedure was able to recover, on average, 63.22% of Cd²⁺, 77.34% of Pb²⁺, and only 1.18% of Cr^total. CASS-BA recovered 56.49% of Cd²⁺, 81.80% of Pb²⁺, and 0.85% of Cr^total. CASS-BB recovered 63.62% of Cd²⁺, 81.17% of Pb²⁺, and 1.14% of Cr^total. In contrast to the adsorbents, AC recovered 71.01% of Cd²⁺, 78.88% of Pb²⁺, and 9.26% of Cr^total.

Considering the excellent recovery rates obtained for Cd²⁺ and Pb²⁺, it is possible to predict the reuse of cassava adsorbents in more than one adsorption cycle. Nevertheless, the same cannot be stated for Cr^total adsorption.

Compared to the obtained results, Manfrin et al., [59], using tobacco-activated carbon, achieved $q_{\text{Langmuir}}=71.42$  mg g^-1 of Pb²⁺, Conradi Junior et al. [3] using tobacco activated carbon reached $q_{\text{Langmuir}}=84.74$  mg g^-1 of Pb²⁺. Thus, superior values than 30.0 mg g^-1 were predicted by using CASS-BK. However, the cassava-based materials are adsorbents without modification, implying lower production costs. The low cost and considerably efficient removal are interesting for adopting this technology in industries wastewater treatments.

Barka et al. [60] reached $q_{\text{Langmuir}}=54.05$  mg g^-1 of Cd²⁺ using Scolymus hispanicus L. adsorbents, against 16.03 mg g^-1 of CASS-BK or [57] that reached $q_{\text{Langmuir}}=9.84$  mg g^-1 of Cr(IV) using cassava sludge-based activated carbon, against 18.38 mg g^-1 using CASS-BK. Thus, cassava-based adsorbents are still competitive with other unmodified and modified cellulose-based materials described in the literature. Therefore, their use in the adsorption of heavy metals is an exciting option.

4. Conclusion

SEM analysis reveals a heterogeneous structure full of cavities. In addition, FTIR before and after adsorption reveals gaps related to missing functional groups, suggesting a significant role of alkenes, carboxylic acid, alcohol, anhydride, and ether. pH_PZC is found at pH 6.02, 6.04, and 6.26, indicating a negatively charged surface above such points, which could potentiate cation adsorption.

In low concentrations of metals, the adsorbents derived from barks, bagasse, and their mixture better perform at pH 7.0 (94.9%) using 16 g L^-1 of adsorbents. Nevertheless, the most cost-benefit dose is found using 8.0 g L^-1, with less than 1% of the maximum efficiency.

The removal of metals reaches equilibrium within 5-10 minutes of contact time with best adjustments for pseudosecond order. The adsorption of metals by cassava adsorbents is better adjusted to the Freundlich model, with significant and critical information provided by Sips, Redlich-Peterson, Temkin, Liu, and Khan models.

Adsorption/desorption studies indicate that cassava adsorbents perform, on average, -10% of the adsorption of metals compared to activated carbon. Nevertheless, factors such as low cost and availability favor the use of such natural materials.