Assessment and modeling of mercury adsorption on carbon-based adsorbents prepared from Jacaranda mimosifolia and guava biomass via pyrolysis and hydrothermal carbonization

The existence of dissolved mercury in water is caused by both natural and anthropogenic pollution sources where the residual discharge from various industries (e.g., mining, energy, battery manufacturing, aerospace, and textiles) plays a significant role in its environmental release (Velempini & Pillay 2019; Prasetya et al. 2020). This heavy metal is toxic and non-biodegradable, and its bioaccumulation in the environment can affect living organisms and pose a risk to public health (Prasetya et al. 2020; Li et al. 2021; He et al. 2023; Shukla et al. 2023; Zhang et al. 2023a). A chronic exposure to mercury has been associated to heart diseases, neurological damage, and cancer in humans (Velempini & Pillay 2019; Prasetya et al. 2020). Toxicological results using laboratory animals have also indicated that the mercury exposure can generate kidney damage, neuropathological effects, decrease in fertility, besides showing carcinogenic activity; while growth retardation has been observed for plants exposed to mercury-polluted water (Syversen & Kaur 2012).

Effluents and wastewater containing mercury can be treated using various physicochemical methods including traditional coagulation and precipitation operations and advanced purification technologies such as membrane-based separation and electrochemical processes (Al-Jaaf et al. 2022; Ali et al. 2022a; Yoo et al. 2023). Particularly, the adsorption-based processes have proved to be effective for the removal of inorganic and organic pollutants from water. For example, Al-Jaaf et al. (2022) showed the benefits of using an eggplant peel-based adsorbent to purify a domestic wastewater. Therefore, the depollution of water containing mercury via adsorption is characterized by its lower cost and higher removal efficiency than those of other treatment technologies (Luo et al. 2020; Duman 2021; Jung et al. 2021). Zeolites, clay, polymers, industrial residues, activated carbons, chars, metal oxides, and their composites have been used as adsorbents to reduce the mercury concentrations in polluted streams (Zhu et al. 2012; Raji & Pakizeh 2014; Ganzagh et al. 2016; Saman et al. 2016; Shen et al. 2018; Prasetya et al. 2020; Jumah et al. 2021; Li et al. 2021; Liu et al. 2022; He et al. 2023). The performance of some of these materials is promising for large-scale water treatment due to their high adsorption capacities, but the main limitation is usually their depollution cost. The operational cost of adsorption systems is mainly related to the production of the solid phase used as the separation medium. Consequently, it is important to identify new feedstocks and low-cost and sustainable preparation routes to improve the economy of water treatment.

Adsorbents prepared by the thermochemical conversion of biomass wastes and subproducts are an interesting alternative for reducing the water purification cost associated with the mercury removal (Jung et al. 2021; Tran et al. 2021; Qin et al. 2023; Zhang et al. 2023a). Carbon-based adsorbents are typically obtained via the thermochemical conversion of biomass, in addition to surface functionalization, to tailor their surface properties. The adsorbent removal performance depends on the precursor feedstock, the synthesis route (e.g., temperature and dwell time), and the chemicals utilized in their functionalization. The quantity and type of adsorbent functional groups (i.e., active sites to bind the water pollutant) are determined by manipulating the synthesis parameters. Therefore, the analysis of the preparation conditions is crucial for establishing the final surface chemistry of the adsorbent.

Pyrolysis and hydrothermal carbonization are reliable biomass conversion routes to obtain materials to reduce the concentration of water pollutants, such as mercury (Correa et al. 2019; Zhang et al. 2023a). They differ in terms of energy consumption, process operating conditions, adsorbent yields, and final product properties (Rattanachueskul et al. 2017; Nguyen et al. 2022). The degradation of the natural polymers contained in biomass determines the development of the porous adsorbent structure and occurs at different temperatures for each preparation route. Therefore, adsorbents prepared from the same feedstock using hydrothermal carbonization and pyrolysis can differ substantially in their surface functionalities and textural properties (Zhang et al. 2023a). The tailoring of adsorbent properties implies the identification of the best biomass conversion conditions to improve the removal of the target adsorbate. Adsorbents for liquid-phase separation can be prepared via the pyrolysis and hydrothermal carbonization of biomass residues such as coconut shells (Jain et al. 2015), sugarcane bagasse (Rattanachueskul et al. 2017), tree leaves (Yang et al. 2019), palm kernel shells (Beri et al. 2021), rice husk (Jung et al. 2021), and corn straw (Zhang et al. 2023a). The porous solids synthesized from the conversion of lignocellulosic biomass residues typically contain oxygenated functional groups. Carboxylic, phenolic, lactonic, ketone, and hydroxyl groups are important for the mercury adsorption on carbon-based surfaces (Jain et al. 2015; Duman 2021; Nguyen et al. 2022). Consequently, it is important to compare the removal performance and properties of the adsorbents obtained using both hydrothermal carbonization and pyrolysis for the available variety of feedstock to identify the best alternative. However, this comparative approach has not been implemented in the preparation of carbon-based adsorbents in previous studies, which usually implies the definition of arbitrary adsorbent preparation conditions that are expected to be far from the optimum separation efficacy–cost balance.

A comparative analysis of the mercury removal properties of adsorbents obtained from pyrolysis and hydrothermal carbonization of Jacaranda mimosifolia fruit and guava (Psidium guajava; Psidium guajava, Linn) seeds is reported in this manuscript. Note that other studies have introduced these biomass wastes as feedstock to obtain activated carbons for the adsorption of pharmaceuticals, heavy metals, phenolic compounds, and dyes (Abdelwahab et al. 2007; Elizalde-González & Hernández-Montoya 2009; Treviño-Cordero et al. 2013; Anisuzzaman et al. 2016; Pezoti et al. 2016; Aly et al. 2019; Ortíz-Gutiérrez et al. 2020; Georgin et al. 2021; Ali et al. 2022b; Pindiga et al. 2022). For example, Elizalde-González & Hernández-Montoya (2009) used the guava seeds as precursor of carbon-based adsorbents for the removal of acid dyes. Anisuzzaman et al. (2016) obtained activated carbon from guava seeds for chlorinated phenol adsorption. This biomass was also employed by Pezoti et al. (2016) as feedstock to prepare activated carbon for amoxicillin uptake. The guava biomass (including leaves, tree bark, and seeds) has also been applied as an adsorbent for the removal of arsenic, fluoride, heavy metals, phenol, and dyes (Domínguez & Serrano 2004; Abdelwahab et al. 2007; Valencia-Leal et al. 2012; Sánchez-Sánchez et al. 2013; Aly et al. 2019; Mohan et al. 2019; Mandal et al. 2020; Ortíz-Gutiérrez et al. 2020; Georgin et al. 2021; Ali et al. 2022b; Behera et al. 2022; Pindiga et al. 2022). Abdelwahab et al. (2007) and Ortíz-Gutiérrez et al. (2020) evaluated the performance of guava seeds for the adsorption of hexavalent chromium. Elizalde-González & Hernández-Montoya (2009) conducted experiments to assess the effectiveness of this residue as an adsorbent of different acid dyes. The guava leaves were employed for the removal of cadmium (Ali et al. 2022b; Pindiga et al. 2022), lead (Pindiga et al. 2022), and dyes (Aly et al. 2019). The adsorption of geogenic pollutants (e.g., arsenic and fluoride) using guava waste was also studied by Mohan et al. (2019), Behera et al. (2022), Sánchez-Sánchez et al. (2013) and Valencia-Leal et al. (2012). On the other hand, Mandal et al. (2020) reported the adsorption of phenol on guava tree bark. Regarding the J. mimosifolia, Treviño-Cordero et al. (2013), Georgin et al. (2021), and Domínguez & Serrano (2004) reported the synthesis of activated carbons using this biomass as precursor for the adsorption of heavy metals, dyes, and ketoprofen.

In this study, the application of the carbon-based materials prepared from these biomass wastes was extended to analyze the mercury removal at pH 4 and 25–40 °C. The main surface chemistry and textural parameters of these adsorbents were also studied and compared. Adsorption equilibrium was investigated using a statistical physics model to characterize the interactions involved in generating the interface responsible for mercury removal. Therefore, the aim of the present study was to assess the surface properties of these alternatives adsorbents to remove mercury from aqueous solutions. The novelty of this study lies on a reliable comparison and analysis of the application of pyrolysis and hydrothermal carbonization of residual biomass precursors to obtain new adsorbents for the removal of mercury as a priority water pollutant.

Deionized water (Mapla), analytical-grade HgCl₂ (Sigma-Aldrich), standard NaOH and HCl solutions (Fisher), and KBr (Sigma-Aldrich) were used in this study. The analytical equipment included: Empyrean (Malvern PANalytical) X-ray diffractometer, Nicolet iS10 (Thermo Scientific) FTIR spectrometer, TM3000 (Hitachi) scanning electron microscope with a coupled energy dispersive system (Nano XFlash, Bruker), ICE 3000 (Thermo Scientific) atomic absorption spectrophotometer and ASAP 2020 (Micromeritics) porosimeter.

Carbon-based adsorbents were obtained from the biomass wastes of J. mimosifolia and P. guajava using pyrolysis and hydrothermal carbonization. These biomass wastes were obtained from local agri-food industries, parks, and gardens. They were then dried, shredded, and screened for the adsorbent synthesis. The pyrolysis of biomass waste was performed in a tubular furnace at 600 °C for 2 h with 10 °C/min heating rate under N₂ atmosphere. Hydrothermal carbonization was conducted in a 100-mL Teflon-lined autoclave reactor using a mass/volume ratio of 1:4 (i.e., 15 g of biomass and 60 mL of deionized water), at 180 °C for 12 h. The results reported in literature and preliminary trials were employed to define these preparation conditions. All adsorbents were washed with deionized water (until a constant pH was obtained from the residual effluent), dried in an oven for 24 h, and sieved to obtain a homogeneous particle size (∼0.67 mm) for the adsorption studies. The samples were labelled as J-PYR, G-PYR, J-HTC, and G-HTC for the solid products obtained from the pyrolysis (PYR) and hydrothermal carbonization (HTC) of jacaranda (J) and guava seed (G) wastes, respectively. A schematic representation illustrating these experimental steps is reported in Supplementary material, Figure S1.

All the adsorbents were characterized to determine their main physicochemical properties. Specifically, the X-ray diffraction (XRD) patterns of these samples were obtained to analyze the crystalline structure and presence of inorganic elements. The adsorbents were analyzed at room temperature with Cu radiation (λ = 1.5406 Å) at 45 kV/40 mA in the angular range of 10°–60° 2θ. The functional groups present on the surfaces of these adsorbents were identified using Fourier transform infrared (FTIR) analysis. FTIR spectra of KBr-adsorbent pellets were recorded (32 scans per sample) in the 4,000–400 cm⁻¹ range with a resolution of 4 cm⁻¹. The elemental composition and morphology of all the adsorbents were determined by Scanning Electron Microscopy (SEM) with Energy Dispersive X-Ray (EDX) analysis. The samples were submitted to high vacuum with an accelerating voltage of 20 kV prior their analysis. Textural properties of adsorbent samples were estimated from N₂ physisorption, which were quantified at 77 K. Finally, the total basic and acidic groups were quantified according to the method described by Boehm (1994) and following the methodology reported by Pawlicka & Doczekalska (2013). The titrations were performed with 0.1 M NaOH and HCl solutions where the corresponding experimental conditions were 1 g/L of adsorbent/solution ratio, 48 h of contact time at 25 °C. Mass balance analysis was carried out with the results from solution titrations to calculate the concentrations of the adsorbent basic and acidic sites. The estimation of pH value of the point of zero charge (pH_pzc) for all the adsorbents was conducted following the methodology described by Milonjić et al. (1975).

As indicated, the mercury adsorption properties of carbon-based adsorbents synthesized from jacaranda biomass were better than those of guava waste-based adsorbents, independent of the chosen preparation route (i.e., hydrothermal carbonization or pyrolysis). This performance may be related to the polymeric composition (i.e., lignin, hemicellulose, and cellulose) of these residues, which affects the adsorbent surface chemistry. Other studies have indicated that jacaranda biomass contains a polymeric fraction of 98% (i.e., cellulose 50%, hemicellulose 21%, and lignin 27%), whereas the polymeric content of guava seed biomass is 77% (i.e., cellulose 61%, hemicellulose 9%, and lignin 7%) (Elizalde-González & Hernández-Montoya 2009; Treviño-Cordero et al. 2013). The type of thermal treatment applied to prepare each adsorbent also affects the presence of surface functionalities, because temperature promotes the degradation of natural polymers contained in the biomass precursor. Hemicellulose starts its thermal decomposition at ∼160 °C, while lignin and cellulose begin to degrade at ∼ 220 and 180 °C, respectively (Heidari et al. 2019). The degradation degree of these polymers depends on the adsorbent preparation process and its conditions (Heidari et al. 2019). Therefore, the type of biomass feedstock is a key parameter affecting the adsorption performance and physicochemical properties of carbon-based materials (Han et al. 2017). Note that the lignin content of the biomass precursor has been associated with a significant impact on the adsorbent surface chemistry, mainly on oxygenated functional groups (Arias-Arias et al. 2017). For instance, previous studies have shown that this polymer undergoes condensation reactions and polymerizes to form the adsorbent structure during hydrothermal carbonization of biomass (Plavniece et al. 2022). The jacaranda waste has a higher lignin content than guava seed residues, thus causing wetting and water diffusion inside the particles to become more constrained (Rodríguez-Correa et al. 2019). This may cause insoluble lignin to generate a highly condensed adsorbent structure (Plavniece et al. 2022), which could limit the mass transfer of dissolved mercury in aqueous solutions, thus affecting the adsorption performance. Consequently, the J-PYR adsorbent sample obtained by pyrolysis exhibited a higher mercury adsorption capacity than J-HTC prepared via hydrothermal carbonization. Wang et al. (2022) also found similar results for the removal of cadmium ions with adsorbents prepared from Napier grass. Table 1 reports the concentrations of the acidic groups found in the tested adsorbent samples. The concentration of acidic sites followed the order: J-HTC > J-PYR ≅G-HTC > G-PYR. The jacaranda-based adsorbents had a higher concentration of oxygenated functionalities (e.g., hydroxyl and carboxyl groups) for mercury binding than other adsorbents (G-HTC and G-PYR). It is also clear that G-HTC outperformed G-PYR because of its higher acidic functional group content. These findings agreed with those reported in the literature regarding the importance of oxygenated functional groups for heavy metal adsorption from water (Arias-Arias et al. 2017; Othmani et al. 2021). On the other hand, it was found that that the leaching of trace elements (e.g., Mg, Fe, Ca, K, Na) from adsorbent surface did not occur in the aqueous solution after mercury removal.

The four adsorbents exhibited different pH_pzc values ranging from 5.08 (J-PYR) to 7.13 (G-PYR), see Table 1. At tested solution pH, all the adsorbents were positively charged, and the mercury adsorption capacity of the tested adsorbents decreased as pH_pzc increased. This trend indicated that electrostatic forces were a minor contributor to the interaction forces during the removal of this metal cation. The corresponding speciation diagram shows that HgCl₂ was mainly present in the aqueous solution at pH 4 (see Supplementary material, Figure S2), thus explaining the negligible effect of the electrostatic forces on the adsorption mechanism. HgCl₂ can interact with the acidic functional groups (mainly hydroxyl and carboxyl groups) of these adsorbents via ligand exchange, as described below (Arias-Arias et al. 2017; Kaveh & Bagherzadeh 2022). All adsorbents showed specific surface areas < 10 m²/g and could be classified as low-porosity materials, see Table 1. These results indicated that the surface area of these adsorbents did not play a relevant role for mercury adsorption, which was confirmed with a statistical analysis. Consequently, it could be expected that the interfacial phenomenon for the separation of mercury from the aqueous solution occurred mainly via the interaction with the functional groups located on the external surface of these adsorbents, with a very limited contribution from intraparticle diffusion.

In addition, van der Waals forces may have been present during the mercury adsorption (Lim et al. 2023).

The mercury adsorption properties of adsorbents prepared from hydrothermal carbonization and pyrolysis of jacaranda and guava seed wastes were analyzed and compared. Mercury adsorption on these adsorbents was endothermic and may involve a multi-molecular interaction mechanism with oxygenated functionalities. van der Waals and ligand exchange are expected to be the main interaction forces for the mercury adsorption mechanism using these adsorbents. The carbon-based materials obtained from jacaranda exhibited the highest mercury adsorption capacities where the biomass pyrolysis allowed to obtain the adsorbent with the best mercury removal. This adsorbent outperformed the adsorption capacities of other activated carbons with higher surface areas. It was expected that the mercury adsorption on these carbon-based materials occurred mainly on their external surface with a limited contribution from intraparticle diffusion. The lignin content of jacaranda waste favored the formation of oxygenated functionalities during biomass thermochemical conversion, especially pyrolysis, under the tested experimental conditions.

The support provided by the MatPore – Porous Materials National Laboratory is acknowledged.

All authors contributed to the study conception and design. V.A.M.-H., F.G.Q.-Á., D.I.M.-C., and A.B.-P. conceptualized the study; V.A.M.-H., F.G.Q.-Á., D.I.M.-C., H.E.R.-Á., I.A.A.-V., and A.B.-P. performed the methodology; F.G.Q.-Á., H.E.R.-Á., D.I.M.-C., H.E.R.-Á., I.A.A.-V., V.J.L.-S., and A.B.-P. investigated the study; F.G.Q.-Á., H.E.R.-Á., I.A.A.-V., and D.I.M.-C. wrote and prepared the original draft; D.I.M.-C. and A.B.-P. wrote, reviewed, and edited the article; D.I.M.-C. supervised the study; D.I.M.-C. did project administration . All the authors have read and agreed to the published version of the manuscript.

This research received no external funding.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.