Activated Carbon from Banana Pseudostem: Multivariate Optimization of Synthesis and Adsorption Study for Phosphorus Removal

This study reports the synthesis of high-surface-area activated carbon from banana pseudostem using zinc chloride activation, with conditions optimized by a central composite rotational design. The optimized material achieved a specific surface area of 2415 m² g⁻¹, significantly exceeding predicted values, highlighting the influence of post-pyrolysis treatment.

Adsorption studies showed that phosphate removal was favored below the point of zero charge (pH 7.30) and followed pseudo-second-order kinetics, reaching equilibrium within 7 h. The Langmuir model best described the adsorption behavior, with a maximum capacity of 11.78 mg g⁻¹. These results demonstrate the potential of banana pseudostem–derived activated carbon for efficient phosphorus removal from water.

The original article

Activated Carbon from Banana Pseudostem: Multivariate Optimization of Synthesis and Adsorption Study for Phosphorus Removal

Moisés de Souza Luz Faria, Tatianny de Araujo Andrade, Renata Pereira Lopes Moreira, Rita de Cássia Superbi de Sousa, and Alisson Carraro Borges*

ACS Omega 2026, 11, 2, 2930–2946

https://doi.org/10.1021/acsomega.5c08820

licensed under CC-BY 4.0

Selected sections from the article follow. Formats and hyperlinks were adapted from the original.

Irregular disposal of wastewater causes negative impacts on the environment, serious damage to human health, and aquatic pollution. Irresponsible human activities in rivers and lakes contribute to their contamination and degradation. Furthermore, the growth of the world population tends to increase the demand for clean water. (1) Fertilization activities, animal residues, and inefficient domestic wastewater treatment enhance the eutrophication levels in water bodies due to their nitrogen and phosphorus content. (2,3) This environmental problem is commonly associated with fast harmful algae growth, which stimulates the surface covering in water systems, leading to oxygen depletion and light obstruction. (3) A variety of alternatives are commonly employed for the treatment of wastewater, such as sedimentation, coagulation, biological techniques, reverse osmosis, and ionic exchange. However, some of these methods may be expensive or cause secondary pollution. (4) Adsorption has been studied in the last decades as a promising technique for the removal of many contaminants, including phosphorus, with different materials being employed as adsorbents. (4−7) This technique promotes advantages compared to other treatment methods, for example, low cost, high effectiveness, easy operation, and fewer byproducts. (8)

Various materials have been developed to enhance the adsorption of contaminants in aqueous solutions. The synthesis of iron-based nanocomposites on activated carbon has shown versatile applications in the removal of toxic metals and dyes, due to the combination of high surface area and active functional groups, acting through electrostatic interactions between the adsorbent and the adsorbate. (9) The use of lignocellulosic residues, such as pea pods and watermelon peels, has also been explored as a sustainable approach for producing chemically activated porous matrices, which are effective at removing Cu²⁺ ions. This performance is enhanced by the incorporation of functional groups and is influenced by factors such as adsorbent dosage, pH, contact time, and the initial concentration of the contaminant. (10,11)

Similar research efforts have focused on creating functional materials, such as three-dimensional electrocatalytic systems based on metal organic frameworks (MOFs), for an efficient degradation of pollutants like norfloxacin. (12) The development of novel nanocomposites has also attracted attention, particularly those combining activated carbon and nanocellulose as the shell with a core formed by cationic metal oxide nanoparticles, enabling the efficient removal of bicarbonate ions. (13)

Phosphorus is an essential nutrient for the agricultural industry. However, its natural reserves are nonrenewable, which highlights the need to develop effective materials capable of capturing its ionic forms from wastewater, since the release of phosphorus into water bodies represents a major environmental concern. (14) Therefore, the development of effective adsorbents, such as activated carbon, is crucial to a sustainable approach. Furthermore, the use of alginate hydrogels has been reported as another efficient adsorbent, since the modification of surface −OH and −COOH groups promotes an enhancement in phosphate removal. (15) Significant progress has also been achieved in biomass-derived sorbents for phosphate remediation as observed in the modification of rice husk biochar with MgAl₂O₄. (16)

In this context, adsorption stands out as an advantageous technique for phosphorus removal due to its efficiency and the possibility of regenerating adsorbents. (17) Phosphate ion removal by adsorption is influenced by the pH value, which determines the electrostatic interactions, by selectivity, since different ions compete for the active sites, and by the initial concentration of phosphorus in the form of phosphate. (18)

Lignocellulosic residues are carbon-rich resources that can be recycled for the synthesis of functional materials. Controlled slow pyrolysis is a process based on the thermal degradation of the source material, resulting in a highly porous product, named biochar. (19) Biochar may enhance the phosphorus adsorption capacity compared to other adsorbents due to its high specific surface area (S_BET), porous structure, and stable chemical properties. (18) The specific surface area of biochars can be improved using different activators, such as physical or chemical reactants. (20) Zinc chloride (ZnCl₂) is a chemical reactant widely used for activation of carbonaceous material for developing high specific surface area structures and porosity formation compared to sodium hydroxide, sulfuric acid, and phosphoric acid. (21−23) Acting as a Lewis acid, this salt induces dehydration of lignocellulosic residues, thereby enhancing the decomposition of their structure during pyrolysis. (24) Innumerable lignocellulosic residues have been studied for activated carbon production, such as sugar cane bagasse, (25) coffee waste, (26) banana peel, (27) cassava starch waste, (28) baobab husk, (29) and banana pseudostem (BP). (7,30,31)

Each ton of banana fruit generates four tons of residues, from which the majority is composed of BP (Musa spp.). This residue may be burned after fruit harvesting or even discharged without proper treatment, promoting the eutrophication of water bodies and negatively impacting the environment. (32) Furthermore, the discharge of BP means the waste of potential and valuable resources. (33) BP is a natural fibrous material with great potential as a precursor of high-quality carbon-rich structures, whose properties may be influenced by different parameters such as activation temperature, pyrolysis time, and activation methods, which affect, for example, the development of surface area. (34) The morphological and structural properties of BP make it a promising precursor for developing porous carbon materials with a high surface area and tunable active sites. Similar approaches have been applied in catalytic degradation systems, where the structural tailoring of nanocomposites such as Sb₂O₃–CuO has been shown to enhance pollutant degradation efficiency. (35)

Although conventional methods for producing activated carbons may yield efficient adsorbents, a systematic study of the production criteria can generate materials with an even higher efficiency by optimizing various factors. On a large scale, this translates into economic and sustainable benefits, as improved process parameters can reduce costs while maintaining an efficient product. (36) The Central Composite Rotational Design (CCRD) is a methodology within Response Surface Methodology (RSM) and represents a simultaneous approach for optimizing multiple factors affecting a dependent variable, making it ideal for evaluating and enhancing adsorption systems. (37) Variables such as activation temperature (AT), pyrolysis time (PT), and impregnation ratio (IR) have been applied in RSM for the synthesis of activated carbon derived from hazelnut shells to evaluate their effects on yield and specific surface area, (36) while the production of hydrogen was also investigated using this methodology, varying the temperature, methane concentration, and catalyst quantity. (38) At the laboratory bench, this also allows a reduction in the number of experiments and enables a multifactor statistical analysis to determine the ideal conditions for all relevant parameters. (39,40) However, many studies still lack a methodology that enables simultaneous optimization of parameters for the synthesis of biochars. In the case of BP, this potential remains underexplored, particularly regarding synthesis under an oxygen-containing atmosphere, selection of an optimal point based on minimal energy consumption, and aiming the material for phosphate adsorption.

Therefore, this study aimed to optimize the production of banana pseudostem activated carbon (BPAC) using ZnCl₂ as a chemical activator and an air atmosphere muffle furnace for phosphorus adsorption. The synthesis of BPAC was optimized, for the first time, using a CCRD methodology within the Response Surface Methodology framework, combining the effects of the activation temperature, ZnCl₂ impregnation ratio, and pyrolysis time on the specific surface area. The optimal conditions were based on the lower levels of the independent factors, maximizing specific surface area values, and a simplified cost-benefit analysis based on the costs of reactant and energy compared to the maximum surface area BPAC obtained by RSM. The optimal BPAC was characterized using techniques such as specific surface area analysis, pore size distribution, X-ray diffraction, scanning electron microscopy, pH of point of zero charge, and ζ-potential. Finally, the kinetic and isotherm batch studies were conducted on the optimal BPAC at 20 °C to determine its equilibrium time and adsorption capacity of phosphate.

2. Materials and Methods

2.5. X-ray Diffraction (XRD)

XRD analysis was performed using a diffractometer (Bruker, D8-Discover) with a Cu-kα radiation source (λ = 1.5418 Å), filtered with a Ni filter. The scans were conducted in the 2θ range from 10 to 70°, with a step size of 0.05°. The samples were fixed in a glass holder with propylene glycol drops added for better dispersion and fixation.

2.6. Scanning Electron Microscopy (SEM)

SEM analysis was performed using a scanning electron microscope (Jeol, JSM-6010LA) with a resolution of 4 nm and an electron beam voltage of 20 kV. The images were acquired in 400× and 15,000× magnifications. Prior to SEM analysis, the samples were coated with a gold layer approximately 12–20 nm thick to enhance conductivity. A complementary analysis of energy-dispersive X-ray Spectroscopy (EDS) was carried out with a silicon drift detector coupled to the microscope with an electron beam voltage of 15 kV and resolution of 133 eV.

2.7. Point of Zero Charge (pHpzc) and ζ-Potential (ζ)

The point of zero charge was performed according to Akkari et al. (2023). (42) Initially, 50 mL of NaCl 0.01 mol L^–1 was disposed in conical flasks (125 mL capacity), varying the pH values from 2.0 to 12.0 in 2 unit increments. For pH adjustment, stock solutions of NaOH 0.1 mol L^–1 and HCl 0.1 mol L^–1 were used. An amount of 0.15 g of BPAC was added to the recipients and then placed in a shaker at 120 min^–1 rotation for 24 h. After that, the pH values were measured again. The pH_pzc was found as the point of intersection between the pH_final × pH_initial curve and the bisector. The assays were conducted in duplicate.

The ζ-potential was determined by using a particle analyzer (Anton Paar, Litesizer 500). Previously, a stock solution of activated carbon 0.5 mg L^–1 was prepared by mixing a certain amount of BPAC with a 0.1 mol L^–1 NaCl solution. Before solid dispersion, BPAC was sieved through a 270-mesh sieve. Then, it was mixed for 24 h using a magnetic stirrer and subsequently subjected to ultrasonic mixing to ensure dispersion for another 2 h. The stock BPAC solution was diluted to a 0.1 mg L^–1 and a 25 mL volume was selected for pH adjustment, varying from 2.0 to 10.0 in 2 units increments, using 0.1 mol L^–1 concentration of HCl or NaOH solutions. After this, an aliquot of approximately 3 mL was put in a cuvette and analyzed in a particle analyzer. This experiment was conducted in duplicate.

3. Results and Discussion

3.4. X-ray Diffraction

Figure 4 shows the diffractograms for the raw banana pseudostem and optimal BPAC. 

ACS Omega 2026, 11, 2, 2930–2946: Figure 4. XRD diffractograms for optimal BPAC (400/2/60) and raw banana pseudostem.

The observed XRD peaks from banana pseudostem at 2θ of 14–17° and 22° correspond to the crystalline structure of cellulose I in planes 1–10 and 200, respectively, with an amorphous valley at 2θ = 18.5°. (82,83) Although additional peaks are observed in the optimal BPAC, none of them are related to zinc oxides (2θ values of 39, 46.3, 49.7, 56.9, 70.5, and 76.8°), indicating an efficient removal of the chemical activator during the washing step. (22)

Peaks of 2θ approximately 2θ at 10.9, 27.9, 31.6, 32.7, 33.4, and 37.6° suggest the formation of lamellar zinc hydroxychlorides, with typical hexagonal nanosheets of simonkolleite (Zn₅(OH)₈Cl₂·H₂O). It is consistent with the high ZnCl₂ impregnation (IR = 2), using a solution of approximately 4.40 mol L^–1 (30 g of ZnCl₂ in 50 mL of deionized water used for impregnate banana pseudostem). (84) Well-defined XRD peaks indicate the development of aligned crystallographic planes, reflecting increased crystallinity, although the majority of the material remains amorphous, as evidenced by diffuse halo around 15° < 2θ < 30°. (82)

Simonkolleite is a zinc chloride monohydrate (layered double hydroxides) with applications in biomedical fields, and water pollutant removal techniques, including adsorption. (85,86) Taglieri et al. (2023), (87) Qu et al. (2023) (88) and Momodu et al. (2015) (89) describe simonkolleite formation by Zn²⁺, OH^–, Cl^–, and H₂O, noting that crystal growth can be slow and sensitive to temperature and pH. In this study, the longer soaking time in HCl solution and the multibatch process may have enhanced the surface development of simonkolleite on optimal BPAC.

Chlorine anions occupy each top corner of the crystal and can be exchanged by other anions via ionic exchange. (90) Co-doping strategies, such as fluorine-chlorine exchange, have been shown to enhance phosphate adsorption by layered double hydroxides, (91) suggesting that the simonkolleite developed in optimal BPAC could improve phosphate adsorption. Simonkolleite is stable in the pH range of 5.5–7.0, and this layered double zinc hydroxide promotes phosphate adsorption through electrostatic attraction and ionic exchange, since the crystal can act as a specific binding site. However, in lower pH values, the removal tends to decrease due to partial dissolution of crystal. (92,93) In this study, the adsorption tests were conducted in a pH range between 4.2 and 4.8, which may result in a partial dissolution of the crystal, indicating the importance of pH control to maintain stability and adsorption efficiency, and minimize potential environmental risks.

3.5. Scanning Electron Microscopy

The SEM images shown in Figure 5 indicate a difference in the surface morphology among the synthesized activated carbon, raw banana pseudostem, and the biochar developed in the same conditions of optimal BPAC, except for zinc chloride addition.

ACS Omega 2026, 11, 2, 2930–2946: Figure 5. SEM images in 400× (a, c, e) and 15,000× (b, d, f). Lines “(a, b)”, “(c, d)”, and “(e, f)” indicate raw banana pseudostem, banana pseudostem biochar synthesized at 400 °C and 60 min without any chemical activator added, and optimal BPAC (400 °C, 60 min, and IR of 2), respectively.

The granules indicated in Figure 5a are bigger than (c) and (e) under the same magnification. The magnified sections (b, d, f) show a more specific morphology. The banana pseudostem (b) has a more uniform surface texture compared to (d), which has a rougher surface morphology, but still smoother than (f). It reveals that pyrolysis by itself may enhance roughness in banana pseudostem due to its thermic decomposition, but the addition of zinc chloride improves the roughness of the final activated carbon. The surface area increased with the implementation of ZnCl₂ as a chemical impregnant. Raw banana pseudostem obtained a specific surface area of 0.904 m² g^–1, while the nonimpregnated biochar (400/0/60) and optimal BPAC (400/2/60) showed 88.058 and 2415 m² g^–1, respectively.

The EDS analysis showed for both pyrolyzed banana pseudostem a high oxygen (21.63 and 24.23%) and carbon contents (78.37 and 71.61%) for biochar without impregnation (400/0/60) and optimal BPAC (400/2/60). For the raw banana pseudostem, the contents of the O and C were close to 44.60 and 49.18%, respectively, and a potassium percentage (K) of 6.22%. The residual missing percentages to complete 100% were not described due to the uncertainty of the equipment. Figure 6 shows spectra obtained for these materials.

ACS Omega 2026, 11, 2, 2930–2946: Figure 6. EDS spectrum for (a) raw banana pseudostem; (b) biochar without impregnation (400/0/60) and (c) optimal BPAC.

The three materials showed a characteristic peak at 2.00 keV, with close highs (225–300 CPS) which can be related to carbon tape used for fixing samples to stub. The spectrum of optimal BPAC included a Zr peak to assess potential cross-contamination from the porcelain crucible after pyrolysis, which may contain this metal in its composition. There was no detection, since the Zr peak could be misinterpreted as uncertainties. The differences obtained in C and O contents demonstrate the effect of volatilization and the enhancement of C in porous matrices from pyrolysis.

4. Conclusions

In the present study, pyrolysis temperature, activation time, and impregnation ratio demonstrated significant effects on surface area development of banana pseudostem activated carbon. The optimal BPAC was obtained under conditions of 400 °C, 60 min, and an IR of 2. The response surface model showed good validation results, with predicted-to-observed ratios ranging from 0.85 to 1.00, using the same method as that applied to the production of carbons from the design. The optimal BPAC exhibited an exceptionally high BET surface area (2415 m² g^–1), surpassing the predicted value. These findings demonstrate the effectiveness of the activation and soaking process in the washing step post-pyrolysis. The prolonged contact between the carbon and the acidic solution may have contributed to the increase in surface area due to an extended interaction time between the acid and BPAC, even at room temperature. The optimal BPAC produced following the criteria of lower energy levels and less reagent consumption, efficiently adsorbed phosphorus compared to other reference biochars, achieving a q_max of approximately 11.8 mg g^–1, with a time of equilibrium reached at 7 h. Overall, the response surface methodology successfully optimized the activation conditions for banana pseudostem activated carbon, confirming the great potential of this lignocellulosic residue as a renewable and efficient precursor for producing high–surface–area carbon materials suitable for adsorption.