Enhanced adsorption of Cu(ii) from wastewater using potassium humate-modified coconut husk biochar

2.5.1 Optimal adsorption conditions

In a batch adsorption experiment, the effects of adsorbent dosage and solution pH on Cu(ii) removal were investigated. To examine the effect of dosage, CB and CMB samples (0.2–4 g) were added separately to 100 mL of a 100 mg·L⁻¹ Cu(ii) solution. The mixtures were shaken at 120 rpm at a constant temperature of 25°C for 540 min. To study the influence of pH, 1 g of CB or CMB was added to 100 mL of a 100 mg·L⁻¹ Cu(ii) solution. The pH of the solution was regulated between 2 and 10 by adding 0.1 mol·L⁻¹ HCl or NaOH. The mixtures were then shaken under identical conditions (25°C, 120 rpm) for 540 min. After the adsorption process, all samples were filtered through a 0.45 µm membrane filter, and the Cu(ii) concentrations were analyzed using a UV–visible spectrophotometer at 600 nm. All experiments were performed in triplicate, and average values were calculated.

In the batch adsorption experiment, the Cu(ii) removal rate R(%) and adsorption capacity q _e (mg·g⁻¹) were calculated using the following equations:

where R is the Cu(ii) removal rate, q _e is the adsorption capacity (mg·g⁻¹), c ₀ and c _e are the initial concentration and equilibrium concentration of Cu(ii), respectively (mg·L⁻¹), V is the volume of the solution (L), and M is the mass of the biochar (g).

2.5.2 Adsorption kinetic experiment

The biochar samples of CB and CMB, each weighing 1 g, were added to 100 mL of a 100 mg·L⁻¹ Cu(ii) solution. The mixtures were shaken at a constant temperature of 25°C with an agitation speed of 120 rpm. Samples were collected at intervals of 30, 60, 120, 180, 300, 420, 540, 720, 900, and 1,620 min. Afterward, the solutions were filtered through a 0.45 µm membrane, and the Cu(ii) concentration in the filtrate was measured using a UV–visible spectrophotometer. Each experiment was performed in triplicate.

The experimental data were analyzed using pseudo-first-order kinetic Eq. 3, pseudo-second-order kinetic Eq. 4, Elovich Eq. 5, and Intra-particle diffusion Eq. 6.

where q _t is the adsorption capacity at time t (mg·g⁻¹), k ₁ and k ₂, respectively, are the pseudo-first-order and pseudo-second-order adsorption rate constants (1/min) and (g·(mg⁻¹·min⁻¹)), t is the adsorption time (min), α is the initial adsorption coefficient (mg·(g⁻¹·min⁻¹)), β is the resolution factor (g·mg⁻¹), K _Ip is the intra-particle diffusion rate constant (mg·(g⁻¹·min^−1/2)) and X _Ip is the intercept (mg·g⁻¹).

2.5.3 Isothermal adsorption experiment

The biochar samples of CB and CMB, each weighing 1 g, were added separately to 100 mL of Cu(ii) solutions at concentrations of 10, 20, 40, 60, 80, 100, 150, 200, 250, 300, 400, and 500 mg·L⁻¹. The mixtures were shaken at a constant temperature of 25°C and an agitation speed of 120 rpm. After 540 min, samples were collected, filtered through a 0.45 µm membrane, and analyzed for Cu(ii) concentration using a UV–visible spectrophotometer. Each experiment was conducted in triplicate.

The equilibrium adsorption behavior of both biochars was evaluated through Langmuir (7), Freundlich (8), and Temkin isothermal models (9)

where q _m is the maximum adsorption capacity (mg·g⁻¹), K _L is the adsorption constant of the Langmuir model (L·g⁻¹), K _F is the adsorption capacity (mg⁽¹⁻ⁿ⁾·L ⁿ /g), and N is the Freundlich model adsorption constant. N = 1, N < 1, and N > 1 indicate adsorption processes of linear, chemical, and physical processes, respectively. R = 8.314 (J·(mol⁻¹·K⁻¹)) is the universal gases. T is the absolute temperature, K _T is the Temkin constant (L·g⁻¹), and b _T is the constant related to the variation in adsorption energy.

2.6.1 Calculation of adsorption column parameters

Total adsorption in column processes (q) is defined as the adsorbate quantity adsorbed from the start to saturation and calculated using

The unit adsorption amount in dynamic adsorption (q ₀) is the saturated adsorption per unit mass of the adsorbent, and computed using

The total removal rate in column adsorption (η) is the ratio of total adsorbed mass to the total mass of adsorbate processed, and determined using

where t _total is the total operating time of the fixed bed (min), c _t is the Cu(ii) concentration at time t (mg·L⁻¹), Q is the influent flow rate (mL·min⁻¹), and m is the adsorbent mass in the column (g).

2.6.2 Column adsorption model analysis

Column adsorption processes were analyzed using the Thomas (13) and Yoon–Nelson models (14). The Thomas model, based on the Langmuir equation, is used frequently to estimate adsorption capacities and rate constants [26]. The Yoon–Nelson model, a classic for column adsorption, is specifically designed for single-component systems [27]

where K _Th is the rate constant of the Thomas model ((mL·min⁻¹)·mg⁻¹), q ₀ is the adsorption capacity (mg·g⁻¹), K _Y is the rate constant of the Yoon–Nelson model (min), and τ is the time required for C _t /C ₀ to reach 0.5 (min).

2.6.3 Column desorption

Desorption experiments were conducted using a 4% HCl solution as the eluent on the adsorption-saturated chromatographic column, at a flow rate of 2 mL·min⁻¹. Effluent samples were periodically collected throughout the desorption process. The effluent Cu(ii) concentration was measured similarly to the batch mode, until Cu(ii) was no longer detectable in the effluent.

3.2.1 Effects of solution pH and biochar dosage on Cu(ii) adsorption performance

Batch experiments were performed to systematically investigate the influences of solution pH and biochar dosage on the adsorption performance of CB and CMB. As illustrated in Figure 2a, solution pH significantly affected adsorption efficiency. When solution pH increased from 2 to 6, the adsorption capacities and removal efficiencies of Cu(ii) for both CB and CMB progressively increased. At strongly acidic conditions, the high concentration of H⁺ extensively protonated biochar surfaces, generating a positive charge [32]. This resulted in strong electrostatic repulsion with positively charged Cu(ii), thus inhibiting metal adsorption [33]. Additionally, excess H⁺ competed directly with Cu(ii) for adsorption sites, further limiting adsorption efficiency [34]. As solution pH rose, decreasing H⁺ concentrations and increasing OH⁻ reduced biochar surface protonation, mitigating electrostatic repulsion. Consequently, electrostatic attraction between negatively charged biochar surfaces and Cu(ii) was enhanced, improving adsorption efficiency [35]. Nevertheless, once the solution pH exceeded 7, the formation of insoluble Cu(ii) hydroxide precipitates significantly affected the adsorption capacity of the biochars. At pH 7, the residual Cu(ii) concentration in solution after precipitation was approximately 0.57 mg/100 mL, whereas the corresponding adsorption capacities of CB and CMB were only 0.21 and 0.23 mg·g⁻¹, respectively. When the pH increased further to 10, the residual Cu(ii) concentration decreased to approximately 0.18 mg/100 mL due to enhanced precipitation, and the adsorption capacities of CB and CMB were reduced accordingly to 0.11 and 0.18 mg·g⁻¹, respectively. These results clearly indicate that at elevated pH values, copper removal is primarily governed by precipitation.

Figure 2

Effects of solution pH (a) and biochar dosage (b) on Cu(ii) removal by CB and CMB (mean ± std). ACB, adsorption capacity of coconut husk biochar (mg·g⁻¹); ACMB, adsorption capacity of modified coconut husk biochar (mg·g⁻¹); RCB, removal rate of coconut husk biochar (%); RCMB, removal rate of modified coconut husk biochar (%); CS, amount of Cu(ii) in 100 mL solution (mg) [conditions: Cu(ii) concentration 100 mg·L⁻¹, adsorption time 540 min, biochar dosage in (a) 10 g·L⁻¹, pH in (b) 6].

Figure 2b illustrates the influence of biochar dosage on Cu(ii) adsorption performance. Adsorbent dosage significantly impacts the adsorption efficiency by determining the availability of active sites at a given initial concentration [36]. As biochar dosage increased from 5 to 10 g·L⁻¹, Cu(ii) removal efficiencies notably improved from 25.67% to 62.58% for CB and from 26.76% to 70.62% for CMB. However, the adsorption capacities per unit mass decreased substantially from 12.06 to 4.50 mg·g⁻¹ for CB and from 13.37 to 5.62 mg·g⁻¹ for CMB. These results indicated that the enhanced Cu(ii) removal efficiencies at higher biochar dosages were primarily attributed to a greater total number of available adsorption sites, rather than improved intrinsic adsorption properties of the biochar itself. When the biochar dosage further increased beyond 10 g·L⁻¹ (10–40 g·L⁻¹), improvements in Cu(ii) removal efficiency became marginal, reaching only 73.62% for CB and 86.52% for CMB. Concurrently, adsorption capacities experienced substantial reductions, decreasing from 4.50 to 2.37 mg·g⁻¹ for CB and from 5.62 to 3.01 mg·g⁻¹ for CMB. This diminishing return at higher dosages can be explained by increased aggregation of biochar particles, potentially causing blockage of adsorption sites and reduced effective surface areas [37]. Additionally, excessive adsorbent dosage could trigger collisions between particles, resulting in the desorption of weakly adsorbed Cu(ii) and subsequently decreasing overall adsorption efficiency[38]. Considering both adsorption effectiveness and economic feasibility, an optimal biochar dosage of 10 g·L⁻¹ was selected for subsequent experiments.

3.2.2 Adsorption kinetics

To further explore the adsorption behavior of CB and CMB, kinetic experiments were conducted, and the results are shown in Figure 3 and Table 1. As depicted in Figure 3a and b, the adsorption capacities of both biochars increased rapidly during the initial stages, owing to the high Cu(ii) concentration and abundance of unoccupied surface binding sites, which generated a strong mass transfer driving force [32]. As the reaction progressed, active sites were gradually occupied, resulting in a decreased adsorption rate until equilibrium was attained.

Figure 3

Adsorption kinetics of Cu(ii) onto CB and CMB (mean ± std). (a) Pseudo-first-order kinetic model, (b) pseudo-second-order kinetic model, (c) Elovich model, and (d) intra-particle diffusion model (conditions: Cu(ii) concentration 100 mg·L⁻¹, pH 6, biochar dosage 10 g·L⁻¹, adsorption time 0–1,620 min).

Nonlinear fitting results revealed that the pseudo-second-order model provided the best correlation with the experimental data, exhibiting higher R ² values (0.9548 for CB and 0.9718 for CMB) and lower RMSE values (1.2500 and 1.9860, respectively) compared to the pseudo-first-order model (R ² = 0.9477 and 0.9539, RMSE = 1.3430 and 2.5450). The corresponding theoretical maximum adsorption capacities were 4.64 mg·g⁻¹ for CB and 6.82 mg·g⁻¹ for CMB, which were in good agreement with the actual measured values of 3.75 and 5.83 mg·g⁻¹, respectively. These results suggest that Cu(ii) adsorption onto both CB and CMB primarily occurred through chemisorption [39].

The Elovich and intra-particle diffusion models were also applied to further elucidate the adsorption mechanisms (Figure 3c and d). The Elovich model exhibited good fitting performance with R ² values of 0.9285 (CB) and 0.9512 (CMB), and RMSE values of 2.6170 (CB) and 1.5720 (CMB), indicating that chemisorption occurred on heterogeneous surfaces [40]. In contrast, the intra-particle diffusion plots displayed multi-linear characteristics and did not pass through the origin, implying that intra-particle diffusion was involved but was not the sole rate-limiting step [35].

As shown in Figure 3d, the Cu(ii) adsorption process could be divided into three distinct phases: an initial rapid phase governed by external (film) diffusion, a slower intermediate phase involving intra-particle diffusion, and a final equilibrium stage. Overall, external diffusion played a significant role in controlling the adsorption rate, while chemisorption remained the dominant adsorption mechanism throughout the process [41].

3.2.3 Isothermal adsorption

Figure 4 and Table 2 present the adsorption isotherm fitting results for Cu(ii) adsorption onto CB and CMB. As shown in Figure 4a and b and Table 2, the Langmuir model provided higher correlation coefficients (R ² = 0.9896 and 0.9826 for CB and CMB, respectively) and lower RMSE values (1.7110 and 2.2050), compared to the Freundlich model (R ² = 0.9690 and 0.9365, RMSE = 3.0920 and 3.5260), suggesting that Cu(ii) adsorption by CB and CMB likely occurred via monolayer adsorption on energetically uniform sites. However, it is important to note that Langmuir adsorption can also remain valid under certain conditions involving multilayer adsorption, especially when the first adsorption layer dominates the overall process. For the Freundlich model, the obtained N values were 0.5975 for CB and 0.5612 for CMB, indicating that Cu(ii) adsorption occurred on heterogeneous surfaces with favorable affinity [42,43].

Figure 4

Adsorption isotherms of Cu(ii) onto CB and CMB (mean ± std). (a) Langmuir and Freundlich models for CB, (b) Langmuir and Freundlich models for CMB, (c) Temkin model for CB, and (d) Temkin model for CMB (conditions: initial Cu(ii) concentration 10–600 mg·L⁻¹, pH 6, biochar dosage 10 g·L⁻¹, adsorption time 540 min).

The fitting results of the Temkin model are shown in Figure 4c and d and Table 2, with R ² values of 0.9536 for CB and 0.9608 for CMB, and corresponding RMSE values of 3.1840 and 3.2940, respectively. Although the Temkin model exhibited slightly lower R ² values and higher RMSE values than the Langmuir model, these values were still within an acceptable range. According to the Temkin model, as surface coverage increased, the adsorption heat gradually decreased, suggesting chemisorption likely dominated Cu(ii) removal.

3.3.1 Adsorption and desorption behaviors of Cu(ii) in fixed-bed column

Figure 5a presents the breakthrough and saturation curves for Cu(ii) adsorption on CB and CMB. During the initial phase, the ratio of outlet to inlet concentration (C _t /C ₀) remained close to zero, indicating a stable removal efficiency above 99%. Over time, the C _t /C ₀ ratios for both materials increased gradually until reaching their respective breakthrough points: 90 min for CB and 300 min for CMB. This phase represents the optimal period for adsorption [44]. CMB maintained a higher removal efficiency for a longer period, exhibiting its enhanced stability and adsorption capacity. After the breakthrough point, the C _t /C ₀ ratio increased sharply, indicating the progressive occupation of active adsorption sites and a significant reduction in adsorption capacity [45]. As adsorption progressed, the C _t /C ₀ ratio for CB and CMB eventually reached their predefined saturation points at 1,080 min and 1,800 min, respectively. Correspondingly, the calculated adsorption capacities were 100.27 mg for CB and 145.20 mg for CMB.

Figure 5

Dynamic adsorption and desorption of Cu(ii) onto CB and CMB (mean ± std): (a) breakthrough and saturation curves for Cu(ii) adsorption onto CB and CMB and (b) dynamic desorption curves of Cu(ii) onto CB and CMB.

Following adsorption, desorption experiments were performed to assess the regeneration potential of CB and CMB, with temporal changes in Cu(ii) concentration in the eluate shown in Figure 5b. Within the initial 10 min, Cu(ii) concentrations rose sharply, peaking at 1,309 mg·L⁻¹ for CB and 1,355 mg·L⁻¹ for CMB, indicating rapid desorption from easily accessible surface adsorption sites. From 10 to 50 min, Cu(ii) concentration declined markedly, signifying the gradual exhaustion of surface-bound Cu(ii) and subsequent desorption from more strongly bound or less accessible sites. Subsequently, from 50 to 420 min, the decline in Cu(ii) concentration slowed notably, reflecting diffusion-limited desorption from deeper internal adsorption sites [44]. After 420 min, the Cu(ii) concentration in the eluates approached zero, signifying completion of the desorption process. It is noteworthy that Cu(ii) concentrations in CMB eluates remained consistently higher compared to CB throughout the desorption process, implying that potassium humate modification not only increased adsorption capacity but also enhanced the subsequent desorption capacity, likely due to additional functional groups and improved accessibility of adsorption sites. The total amounts of Cu(ii) desorbed were 83.27 mg for CB and 91.35 mg for CMB, corresponding to desorption efficiencies of 83% and 63%, respectively. The relatively lower desorption efficiency of CMB is likely due to the stronger chemical binding mechanisms introduced by potassium humate modification, which rendered Cu(ii) more difficult to desorb [33].

In this study, 4% HCl was employed for Cu(ii) desorption due to its high efficiency. However, its potential environmental impact should not be overlooked, and greener, biodegradable eluents will be actively explored in future work. Additionally, further studies will assess CMB’s selectivity, stability, and reusability under realistic wastewater conditions through competitive ion tests and multi-cycle adsorption–desorption experiments.

3.3.2 Fixed-bed kinetic modeling

The experimental data for the column adsorption of Cu(ii) by CB and CMB were analyzed using the Thomas and Yoon–Nelson models (Figure 6 and Table 3). The fitted results showed that the correlation coefficients (R ²) for CB and CMB in the Thomas model were 0.9164 and 0.9352, with corresponding RMSE values of 0.7190 and 0.5560 (Figure 6a). These results indicate a strong fit and reliable prediction of their column adsorption behavior [26].

Figure 6

Thomas (a) and Yoon-Nelson (b) model fitting curves for dynamic adsorption of Cu(ii) onto CB and CMB (mean ± std).

For the Yoon–Nelson model, the R ² values were 0.9012 for CB and 0.9250 for CMB, with corresponding RMSE values of 0.7530 and 0.5920, demonstrating a reliable representation of the adsorption process under different conditions (Figure 6b). The increased number of adsorption sites in CMB extended the breakthrough time from 649 min (CB) to 1,054 min (CMB), indicating a significant improvement in adsorption capacity as a result of the modification. Moreover, an increase in the Yoon–Nelson rate constant (K _Y ) shortened the corresponding adsorption time, resulting in a steeper breakthrough curve and an earlier saturation point [46]. The fitted curves showed that the K _Y value for Cu(ii) adsorption on CB was higher than that on CMB, suggesting that CB reached saturation more quickly.

3.4.1 SEM-EDS analysis

The microscopic surface morphologies of the adsorbent materials were observed by SEM [47]. Initially, the surface of CB appeared relatively uniform and smooth, as shown in Figure 7a and b. After modification with potassium humate, the morphology significantly changed, exhibiting an uneven, aggregated, film-like coating, along with numerous small spherical particles on the surface of CMB (Figure 7d and e). These observations clearly confirmed the successful immobilization of potassium humate onto the biochar surface. After Cu(ii) adsorption, SEM images of CMB-Cu (Figure 7h and i) showed that the originally porous structure was partially filled or blocked, indicating that Cu(ii) were successfully adsorbed onto the biochar surface and into its pores.

Figure 7

SEM images and EDS elemental mapping of CB, CMB, and CMB-Cu. (a) and (b) SEM images of CB, (c) EDS mapping of CB, (d) and (e) SEM images of CMB, (f) EDS mapping of CMB, (g) and (h) SEM images of CMB-Cu, and (i) EDS mapping of CMB-Cu.

EDS analysis provided further evidence supporting these findings. As shown in Figure 7c, f and j, CB primarily contained C and O elements. After modification, a new potassium (K) peak appeared in the EDS spectrum of CMB, verifying the presence of potassium humate. Following Cu(ii) adsorption, the EDS spectrum of CMB-Cu displayed a distinct Cu peak, and corresponding elemental composition changes were observed: C content increased from 50.41% to 68.50%, O content decreased from 49.35% to 28.08%, K content dropped from 0.24% to 0.00%, while Cu content rose from 0.00% to 3.87%. These compositional changes confirmed the successful adsorption of Cu(ii) onto CMB, suggesting interactions between Cu(ii) and the functional groups introduced by potassium humate. Additionally, the observed decrease in K content indicated that ion exchange played a role during adsorption. EDS elemental mapping further demonstrated uniform distribution of these elements, supporting homogeneous adsorption on the biochar surface, which was consistent with the monolayer adsorption behavior described by the Langmuir isotherm model.

3.4.2 FTIR analysis

FTIR analysis was conducted to investigate changes in surface functional groups of CB, CMB, and their forms after Cu(ii) adsorption (CB-Cu, CMB-Cu). As shown in Figure S2, both CB and CMB exhibited similar characteristic peaks. However, notable differences were observed in peak intensities and positions. Compared to CB, CMB exhibited significantly enhanced peak intensities at approximately 3,448 cm⁻¹ (O–H stretching vibrations) [48], 2,827 cm⁻¹ (C–H stretching vibrations of –CH and –CH₂ groups) [49], and 1,608 cm⁻¹ (C═O stretching vibrations from carboxyl groups or aromatic structures) [50]. These changes indicated that potassium humate modification successfully introduced additional oxygen-containing functional groups onto the biochar surface.

After Cu(ii) adsorption, noticeable changes were observed in the FTIR spectra of CB and CMB (denoted as CB-Cu and CMB-Cu, respectively), including reduced peak intensities and slight shifts in key functional group bands. In the case of CMB, the O–H stretching vibration band shifted from 3,448 to 3,435 cm⁻¹ and showed a marked decrease in intensity, indicating that hydroxyl groups were actively involved in coordinating with Cu(ii). Likewise, the absorption band corresponding to C═O stretching in carboxyl groups shifted from 1,608 to 1,614 cm⁻¹, accompanied by a weakening of peak intensity. This suggests that carboxyl groups participated in Cu(ii) binding through deprotonation and complexation processes [51].

Compared with CB, these spectral changes were more prominent in CMB, highlighting the role of additional oxygen-containing functional groups introduced through potassium humate modification. The more substantial peak shifts and reductions in intensity confirmed that the enriched hydroxyl and carboxyl functionalities in CMB played a critical role in enhancing Cu(ii) adsorption, primarily via chemisorption mechanisms.

3.4.3 BET analysis

The textural characteristics of the biochars, determined through BET analysis (Table 4 and Figure S3), showed that CMB possessed a significantly improved surface area and pore structure compared to CB. Specifically, the specific surface area and total pore volume of CMB were 6.3895 m²·g⁻¹ and 0.002364 cm³·g⁻¹, respectively – substantially higher than those of CB (3.5221 m²·g⁻¹ and 0.000795 cm³·g⁻¹). Furthermore, CMB exhibited a more developed micro and mesoporous network, which provided more accessible adsorption sites and promoted faster Cu(ii) diffusion into the internal structure [52]. These improvements were consistent with the results of the Langmuir isotherm analysis, where CMB demonstrated higher maximum adsorption capacity (q _m) and equilibrium affinity constant (K _L ) than CB. This indicates a greater number of active binding sites and stronger interactions between Cu(ii) and the modified surface.

Overall, the increased specific surface area and enhanced pore volume contributed directly to the improved Cu(ii) adsorption performance of CMB. The BET analysis thus supports the conclusion that potassium humate modification effectively improved the structural properties of the biochar, enhancing its adsorption capacity for Cu(ii).

3.4.4 Boehm titration analysis

The surface acidic functional groups (carboxyl, lactonic, and phenolic hydroxyl groups) on CB and CMB were quantitatively determined using the Boehm titration method. As shown in Table 5, CMB exhibited a significantly higher total concentration of acidic functional groups than CB, particularly a marked increase in carboxyl groups (1.50 mmol·g⁻¹ for CMB vs 0.35 mmol·g⁻¹ for CB). Generally, carboxyl groups are among the most acidic oxygen-containing functional groups on biochar surfaces and play a crucial role in the adsorption of heavy metals through ion exchange or complexation mechanisms [53]. Thus, the substantial increase in carboxyl groups on CMB enhanced its capacity to effectively bind Cu(ii), primarily through the formation of surface complexation reactions. Moreover, the increased concentrations of lactonic and phenolic hydroxyl groups in CMB (0.35 and 0.90 mmol·g⁻¹, respectively, compared to 0.12 and 0.45 mmol·g⁻¹ in CB) further enhanced the adsorption capacity by providing additional binding sites for Cu(ii). These conclusions were consistent with the FTIR analysis. Collectively, the results of Boehm titration analyses clearly indicated that the potassium humate modification significantly improved Cu(ii) adsorption through enhanced chemisorption involving carboxyl and hydroxyl functional groups on the biochar surface.

In summary, although the BET surface area of CMB was relatively low (6.4 m²·g⁻¹), the potassium humate immobilized on its surface chemically interacted with Cu(ii) in aqueous solution, thereby enhancing the adsorption performance primarily through surface complexation and ion exchange mechanisms.