Surface modification of spent coffee grounds using phosphoric acid for enhancement of methylene blue adsorption from aqueous solution

With the development of various industries, the dye market has also grown tremendously, which is directly related to water pollution. Dyes account for the largest proportion of water pollutants. Among the various dyes, one notable organic dye is methylene blue (MB), which is widely applied in textiles, chemical indicators, dyeing, printing, pesticides, paper coatings, cosmetics, and pharmaceuticals (Choi & Yu 2019). MB is a heterocyclic aromatic compound and exists in the form of three water molecules bonded to one MB molecule in aqueous solution (Han et al. 2020). In addition, MB contains an aromatic ring; it is toxic and carcinogenic (Lee & Choi 2021). Therefore the removal of MB from an aqueous environment is necessary to maintain water quality and a green environment.

Lignocellulose-based activated carbon is also being studied to adsorb harmful substances by reforming activated carbon with clay (Khnifira et al. 2022), metal (Arora et al. 2019; Soni et al. 2020a; Karami et al. 2022), and metal oxide (Patil & Shrivastava 2016; Soni et al. 2020b; Kumar & Kumar 2022). However, most are produced through two main processes: carbonization or pyrolysis and activation (physical or chemical). Physical pretreatment includes mechanical destruction, boiling, heating, autoclaving, freeze drying, or vacuuming (Basu et al. 2018). However, the adsorbent through physical pretreatment has disadvantages in that the treatment efficiency of harmful substances in aqueous solution is low and selective adsorption is difficult (Han et al. 2020). Therefore, recently, many researchers have activated the surface of the adsorbent by attaching a functional group that is beneficial to the adsorption of harmful substances using chemicals such as acids and bases. As a result, it has been possible to selectively adsorb harmful substances, and the adsorption efficiencies have been reported to increase (Kouzbour et al. 2019; Lee & Choi 2021; Misran et al. 2022). Choi et al. (2021) reported that the carboxyl group (–COOH) of the orange peel adsorbent modified with sulfonic acid was increased, thereby improving the Cu(II) adsorption efficiency. Also, Chen et al. (2021) reported that introducing a carboxylic acid functional group (–COOH) on the surface of the spent coffee grounds (SCG) provides a higher binding capacity to adsorb heavy metal ions. This is because acidic functional groups on the surface of the adsorbent such as phenol, carboxyl, hydroxyl and carbonyl functional groups of the SCG adsorbent modified with sulfonic acid were removed by binding with heavy metal ions (Leng et al. 2019; Elgarahy et al. 2021).

One of the promising lignocellulose-based activated carbon precursors is coffee grounds. Coffee is loved by many people around the world, the second most consumed beverage in the world after water; it is sold with an average of billions of cups per day, and it is known that 6 million tons of soluble coffee are produced annually (Choi et al. 2021). Due to this, a large amount of SCG is produced, and it is expected that the amount of SCG produced every year will steadily increase with the growth of the coffee industry. SCG is a lignocellulosic material composed of lignin (27.5%), cellulose (45.6%), and hemicellulose (21.3%) (Choi et al. 2021; Shi et al. 2021). Adsorption of hazardous substances using SCG has a very useful value in terms of recycling waste. So far, many studies (Choi et al. 2021; Elgarahy et al. 2021; Qiao et al. 2021) have been published on the adsorption of various heavy metals using SCG, but there have been very few papers on the phosphorylation of SCG to remove MB from aqueous solution and analyze the removal efficiency. Phosphoric acid in aqueous solution, on the one hand, promotes the bond cleavage reaction and, on the other hand, promotes crosslinking through cyclization and condensation to form a bond layer such as phosphate and polyphosphate ester, which can protect the internal pore structure (Leng et al. 2019; Guediri et al. 2020; Elgarahy et al. 2021). These reactions can act as excellent MB adsorbents in aqueous solutions. According to previous studies (Xu et al. 2014; Basu et al. 2018; Han et al. 2020; Zeng et al. 2021), phosphoric acid-activated adsorbents can be recycled and have various advantages such as low toxicity. Therefore, in this study, to increase the removal efficiency of MB in aqueous solution, SCG was modified and activated using phosphoric acid. For the optimization of MB removal efficiency, various parameters (initial pH, contact time, initial concentration of MB, capacity of adsorbent) were evaluated. The experimental data were analyzed by adsorption isotherms, and the properties and adsorption kinetics of the adsorbent were used to explain the adsorption mechanism.

SCG was collected at a coffee shop in Gangneung, and ground with a particle size of less than 40–60 mesh (0.25–0.4 mm) were recovered and washed several times with distilled water to remove contaminants. The washed SCG was dried at 75 ± 2 °C for 24 hours to evaporate moisture. The pretreated SCG was stored in a desiccator for activation with phosphoric acid.

The method for activating SCG into porous carbon using phosphoric acid (H₃PO₄) is as follows. First, 200 mL of phosphoric acid (85%) and 75 g of phosphorus pentoxide were mixed and completely dissolved while heating and stirring and then cooled at room temperature. A 100 g sample of dried SCG (40–60 mesh) was added to the prepared phosphoric acid solution (300 mL), and a phosphorylation reaction was induced while stirring at 150 rpm for 5 hours using a cylindrical reactor (3 L). The mixed sample was left at room temperature for 4 hours to allow the reaction to occur completely, then the pH was adjusted to neutral by adding 1 N sodium hydroxide, washed thoroughly with distilled water, and dried for 24 hours in a dryer. Phosphorylated SCG (PSCG) was stored in a desiccator for use in experiments.

Methylene blue (MB; C₁₆H₁₈ClN₃S·3H₂O, molar mass: 319.85 g/mol, Sigma-Aldrich, USA) is a basic dye. The hydrate is blue and exists in the form of three water molecules bonded to each MB molecule. MB was used in powder form. After preparing the 1,000 mg/L concentration, it was then diluted with distilled water to prepare a solution of the required concentration.

The surface images were taken of PSCG and SCG using scanning electron microscopy (SEM, JSM-IT500, JEOL Ltd Japan) and the chemical composition of the adsorbent were analyzed using energy-dispersive X-ray spectroscopy (EDS, JSM-IT500, JEOL Ltd Japan). Fourier transform infrared (FT-IR) spectroscopy (Perkin Elmer, FT-IR 1760X, USA) was applied in a scan range of 400–4,000 cm⁻¹ to investigate the presence of functional groups of SCG and PSCG. Brunauer–Emmett–Teller (BET) surface area was examined using a BET Surface Analyzer (Quantachrome Instruments version 11.03) via adsorption/desorption isotherms of N₂ performed at 77 K. After the adsorption process, the morphology, functional groups and BET surface area of the dried PSCG were also examined to compare before and after adsorption. pHpzc was analyzed according to the method reported in a previously published study (Choi et al. 2021). The amount of adsorbent was measured using an electronic balance (XP26, Mettler Toledo, Swiss), and the pH was measured using a pH meter (SevenGO pro, Mettler Toledo).

Adsorption capacity experiments according to various initial concentrations were analyzed using adsorption kinetics (pseudo-first-order (PFO) and pseudo-second-order (PSO)) and adsorption isotherms (Langmuir, Freundlich and Temkin). Table 1 summarizes the mathematical equations and linearized forms of adsorption kinetics and adsorption isothermal equations.

To test the reusability of PSCG, resorption experiments were performed six times using the same PSCG. As initial solutions, 50 mg/L MB solution and 1 g/L PSCG were prepared. The initial solution was reacted with PSCG for 120 minutes, sampled, measured with a UV spectrophotometer (Shimadzu 1800, Japan), and the removal efficiency was calculated using Equations (1) and (2). After the first reaction experiment, PSCG was dried in an oven at 75 °C for 3 hours, and the same conditions as the initial experiment were used. The PSCG adsorbent reuse experiment was repeated with the same procedure as the first up to six times without a desorption step.

The surface morphology of SCG and PSCG was obtained using SEM (Figure 1(a)). Compared with SCG, PSCG was detected to have increased pores, with the rough surface while the size and shape were heterogeneously changed. This surface morphology provides a great advantage for adsorption, because it can provide more adsorption sites for binding the adsorbate. According to the EDS analysis, it was revealed that the carbon content decreased, but the oxygen and phosphorus contents clearly increased (Figure 1(b)). As the adsorption of MB ions by the modified PSCG was achieved by a surface complexation mechanism between MB ions and oxygen-containing functional groups including phosphorus, an increase in oxygen content and an increase in phosphorus have a favorable effect on MB adsorption.

The FT-IR spectra of SCG and PSCG showed broadening and strengthening of certain bands according to the activation, but there was no significant difference in the basic spectra (Figure 1(c)). SCG contains OH aromatics (950–1,300 cm⁻¹), CH-stretching (1,400–1,450 cm⁻¹), C = O carboxylic acids (1,630–1,750 cm⁻¹), CH-stretching of alkane groups (2,800–2,950 cm⁻¹) and carboxyl, phenol or hydrogen-bonded hydroxyl groups in alcohols (3,100–3,600 cm⁻¹).

Conversely, PSCG contains N-containing bioligands (400–500 cm⁻¹), C = C aromatic compounds (625–730 cm⁻¹), OH aromatics and N-H groups (805–910 cm⁻¹), P = O groups, and carboxyl groups (1,007–1,110 cm⁻¹) and (RO)₃P = O (1,300–1,500 cm⁻¹) were either newly created or broadened. In particular, the 1,017–1,100 cm⁻¹ and 1,300–1,500 cm⁻¹ peaks were deeper and broader than that of SCG because phosphate groups were attached there. These peaks are characteristic of phosphorus and phosphorus–carbon compounds such as phosphates or hydrogen bonds of polyphosphates (P = O groups, P-O-C (aromatic) bonds and P = OOH etc.) present in phosphoric acid-activated carbon (Leng et al. 2019; Choi et al. 2021; Yang et al. 2020). In addition, the peak at 1,000–1,110 cm⁻¹ may be due to the symmetrical vibrations of the P + O– of the acidic phosphate ester and the P–O–P of the polyphosphate chain (Chen et al. 2021; Elgarahy et al. 2021; Qiao et al. 2021). Summarizing the above results, as a result of reforming SCG with phosphoric acid, the number of functional groups increased, and the functional groups required for MB adsorption became wider or deeper. This is because the aliphatic residue contained in SCG was destroyed due to phosphorylation of SCG. However, the 3,100–3,700 peak was decreased in PSCG because the carbon content decreased after phosphoric acid activation of SCG (see Figure 1(b)).

Moreover, the FT-IR spectra for PSCG after MB adsorption showed some functional groups derived from MB. Some of the new peaks are the chromophores NH (996 cm⁻¹), C-S (758 cm⁻¹), and the broad bands of C = C, C = O, C = S, C = N from MB were indicated between 1,240–1,780 cm⁻¹. The presence of these functional groups can be used as evidence that MB is adsorbed to PSCG. The decrease in peak height and shift in peak position also confirm that the adsorbent and adsorbate have a strong interaction. Phosphorylation of SCG increases the surface area and porosity of the adsorbent as well as the presence of functional groups such as aromatic rings, –C = O, –C–O–C–, –OH, –C = N, –P = O, –P–O–C, and P = OOH play important roles in increasing MB adsorption capacity in aqueous solution. The surface of PSCG modified with phosphoric acid contains a significant amount of phosphorus-compounds. IR spectra of SCG-derived carbons show multiple functions observable for other carbons obtained by phosphoric acid activation of lignocellulose precursors.

Table 2 summarizes the component analysis results of SCG and PSCG. As the O/C ratio increases, the anionic properties of the bio-adsorbent were strengthened due to the increase in the carboxylate groups on the surface of the adsorbent, which tends to increase the adsorption capacity of cationic substances (Choi & Yu 2019; Choi et al. 2021). After phosphorylation of SCG, the O/C ratio increased from 0.49 to 0.97. This proves that the carboxyl group inside the adsorbent increased after the phosphorylation of SCG.

In SCG, 86.7% of micropores and 12.9% of mesopores had more ultrafine pores than mesopores. However, after activation with phosphoric acid, the range of mesopores increased to 96.9%, confirming that the overall pore size of the adsorbent increased. In the activation step of SCG using phosphoric acid, the high molecular compounds contained in SCG were destroyed into small molecules, and the pore size increased. As previously reported (Qiao et al. 2021) for the activation process of other lignocellulosic precursors, it is hypothesized that H₃PO₄ reacting within the inner cellulosic structure induces depolymerization leading to enhancement of the pore volume, leading to overall volume expansion. The size of PSCG pores after MB adsorption showed little change in the size of 2–50 nm. However, it decreased below 2 nm and increased above 50 nm, resulting in an overall increase in the pore size of PSCG. The pore size of PSCG increased above 50 nm, resulting in an overall increase in PSCG pores. This is because the pores of PSCG were filled with MB dye molecules after MB adsorption and the pore size increased.

The results of BET measurement (Table 2) show the characteristics of SCG and PSCG. The surface area was significantly increased in PSCG compared to SCG. The activation process of SCG using phosphoric acid successfully enlarged the pores of SCG. This is considered to increase the surface area of PSCG as long-chain polymers such as aliphatic are destroyed by phosphoric acid, as observed in FT-IR analysis. The high surface area of PSCG provides sufficient active sites for MB dye adsorption. However, the surface area of PSCG decreased after adsorption because MB dye molecules were adsorbed on the PSCG surface and covered the pores. Due to this, the pore volume of PSCG was also reduced after MB adsorption.

The pH, together with the size and temperature of the adsorbent pore size and the amount of adsorbent, is a major factor affecting bio-adsorption (Choi et al. 2021). This is because the optimal pH value for a specific biosorbent process depends on the zero point charge value (pHpzc) of the biosorbent and the target metal ion (Basu et al. 2018; Kouzbour et al. 2019). pHpzc is the pH (of the solution) value at which the surface charge of the biosorbent becomes zero (Choi & Yu 2019; Lee & Choi 2021). Proton binding by all carbons exhibits positive (proton adsorption) and negative (proton emission) moieties (Sych et al. 2012; Choi et al. 2021). At a pH above pHpzc, the surface of the adsorbent becomes negatively charged, and the electrostatic interaction with the positively charged adsorbent material is strong, increasing the adsorption efficiency (Xu et al. 2014; Han et al. 2020). When phosphoric acid is added to SCG, carbon with an acidic surface is formed, which shows that the pHpzc significantly decreases from 6.43 to 3.96 in the acidic region (Figure 2(a)). It can be seen that the acidic group was strongly attached to the surface of PSCG. In addition, it means that the pH range that can accommodate the adsorption of MB ions in aqueous solution was widened, which may lead to an increase in adsorption capacity.

In order to examine the adsorption efficiency of MB according to the amount of SCG and PSCG, the experiment was performed at the adsorbent rate of 0.5–3 g/L, the concentration of MB at 20 mg/L, pH 7, and 25 °C. The results are shown in Figure 2(b). In both SCG and PSCG, as the amount of adsorbent increased, the time to reach the maximum adsorption amount shortened, and the total adsorption amount of MB in aqueous solution increased. The removal of MB from the aqueous solution proceeded very rapidly during the first 10 minutes, and then the removal rate was slowed until equilibrium was reached. The rapid removal of MB at the beginning of the reaction is probably due to a large number of vacant sites available for adsorption. With 1 g/L of PSCG adsorbent, 99% removal is possible within 20 minutes. Also, when using PSCG, MB was completely removed at 60 min by all adsorbent doses experimented in this study. The adsorption efficiency of MB for PSCG was about 2.5–3.5 times higher than that of SCG, and it can be confirmed that the maximum adsorption was reached within a short time. This result is thought to be because, as already mentioned in the analysis of the physical property and the surface morphology of the adsorbent, the introduction of various functional groups and phosphate groups on the surface of the adsorbent. Also, the increase in the surface area of ​​the adsorbent after phosphorylation gave favorable adsorption of MB. These results are consistent with the research results of many researchers who reformed agricultural waste to adsorb heavy metals in aqueous solutions (Yang et al. 2020; Liu et al. 2021b; Qiao et al. 2021; Misran et al. 2022).

In general, as the amount of biosorbent increases, the adsorption capacity of heavy metals increases (Zhuang et al. 2020). This is because the number of active sites capable of adsorbing the adsorbate increases as the amount of adsorbent increases (Leng et al. 2019; Zhang et al. 2021). However, according to the results of several previous experiments (Basu et al. 2018; Yang et al. 2020; Choi et al. 2021; Elgarahy et al. 2021), it was observed that, after reaching the maximum adsorption efficiency, the adsorption capacity did not increase but even started to decrease even when the amount of biosorbent was increased. The reason is that if too much biosorbent is administered, the biosorbent particles are agglomerated. As a result, the active sites of the adsorbent overlap each other, and mass transfer is blocked. To solve these problems, increasing the stirring speed of the aqueous solution helps to overcome the mass transfer resistance, but it is necessary to optimize the stirring speed because if the stirring speed is too high, fragmentation of the biosorbent may occur.

According to the experimental results of the amount of adsorbent, it was found that the adsorption efficiency of MB using PSCG was significantly higher than that of SCG. The efficiency at which the various initial concentration of MB was adsorbed onto PSCG was specifically evaluated with 1 g of adsorbent. As the concentration of the MB concentration in the solution increased, the removal efficiency decreased, but the adsorption capacity increased. The adsorption of MB onto PSCG at an initial concentration of less than 150 mg/L showed a high removal efficiency of over 90%. Even at a concentration of 200 mg/L, a removal efficiency of 83.5% and a maximum adsorption capacity of 167.52 mg/g were obtained. These results are shown in Figure 2(c), which is used to obtain a suitable isothermal adsorption model.

To examine the effect of pH on the adsorption efficiency of MB, the pH was controlled from 1 to 10 by a 1 g/L adsorbent dose at 25 ± 1 °C. MB removal in aqueous solutions increased rapidly for PSCG from pH 4 and SCG from pH 7 (Figure 2(d)). This phenomenon was related to the value of pHpzc mentioned above. In addition, as the pH increased, the sulfonic acid, carboxyl, and phenolic hydroxyl groups of the PSCG and SCG adsorbents were deprotonated to form –P = O, –P–O–C, P = OOH, R–COO–, and R–O– groups. Due to this, the surfaces of the PSCG and SCG adsorbents became negatively charged, improving the electrostatic attraction between PSCG and SCG adsorbents and MB ions in aqueous solutions.

The linear model has a simple interpretation of the parameter and can relatively easily calculate correlation coefficients (R²) compared with nonlinear model (Choi & Yu 2018). Therefore, in statistical modeling, which emphasizes the interpretation of the model, linear models are mainly used. In this paper, three linear kinetic models (pseudo-first-order (PFO), pseudo-second-order (PSO) and intraparticle diffusion) were applied to predict the kinetic involved during the present adsorption process. Analysis of the adsorption kinetic model can help to understand the physicochemical interaction between the MB and PSCG, the mass transport, and the adsorption rate, which are necessary to determine the optimal conditions for the adsorption process. In PFO, the adsorption site occupancy is directly proportional to the number of vacant adsorption sites (Amel et al. 2012; Khnifira et al. 2022). However, in PSO, the adsorption of harmful substances is considered a function of the number of free active sites and the number of ionized substances in the solution (Gomora-Hernandez et al. 2020). The analysis results of adsorption kinetics are represented in Table 3, and shown Figure 3(a) and 3(b). To obtain an appropriate kinetic model, the adsorption capacity was be analyzed using different concentrations of MB solutions with different contact times.

The adsorption kinetics showed that the values of q_e,cal of PSO were closer to those of q_e,exp. The value of the correlation coefficient (R²) of PSO was higher, and the value of χ² was lower than that of PFO. Therefore, the MB adsorption process in aqueous solution using PSCG was closer to PSO than to PFO. Moreover, the adsorption rate constant k₂ was higher than that of k₁, confirming that the adsorption rate-determining steps were diffusion and sorption of MB ions into the pores of PSCG. Many studies (Choi & Yu 2019; Leng et al. 2019; Choi et al. 2021; Elgarahy et al. 2021; Zhang et al. 2021) have reported that MB adsorption follows the PSO kinetic model. Moreover, the values of k₁ and k₂ of PFO and PSO decreased as the MB concentration increased, and the adsorption rate of MB onto PSCG decreased as the MB concentration increased.

The plot of q_t versus t^1/2 for MB adsorption onto PSCG (Figure 3(c)) did not pass through the origin, therefore intraparticle diffusion was not the only rate-limiting step. There were three adsorption processes (film diffusion, pore diffusion and intraparticle transport) of adsorbate onto adsorbent (Choi & Yu 2018). The slowest of the three steps controls the overall rate of the process. As found in this study, pore diffusion and intraparticle diffusion are often rate limiting in a batch experiment. The calculated value of k_id and C from the slopes of the plots increased when increasing the MB concentration in the aqueous solution. The larger the value of C representing the thickness of the boundary layer, the greater the contribution of surface sorption in the rate-controlling step.

According to the Langmuir model, adsorption can only occur in a monolayer manner, in which only a single layer of molecules is attached to the surface of the biosorbent (Choi et al. 2021; Zhang et al. 2021). The Freundlich model has a heterogeneous surface of the biosorbent and the adsorption is expressed as a multilayer with different intermolecular interactions (Choi & Yu 2019). It is particularly suitable for application when the concentration of the adsorbent in the aqueous solution is low (Amel et al. 2012; Lee & Choi 2021). The Temkin isotherm is used to evaluate the relationship between the biosorbent surface and the heat of adsorption of all molecules, showing that the heat of adsorption of all molecules decreases linearly with increasing biosorbent surface coverage (Basu et al. 2018; Leng et al. 2019). The R² of the Freundlich model is lower than that of the Langmuir model. Therefore, the interaction between PSCG and MB is performed as a monolayer adsorption that attaches to the biosorbent surface, with the main adsorption as hypothesized by the Langmuir model. Similar findings have been mentioned in several previous studies (Zieliński et al. 2019; Zhuang et al. 2020; Choi et al. 2021; Zhang et al. 2021). However, R² of both the Freundlich and Langmuir models was found to be higher than 0.9 (Table 3), so the adsorption of MB using PSCG can be explained as monolayer and multilayer adsorption, applying to both models. Freundlich's K_F indicates that the larger the value, the better the adsorption ability, and the 1/n value is 0.1–0.5 (excellent adsorption process), 1/n = 0.5–1 (the process is easily adsorbed), and 1/n ≥1 (the process is difficult to adsorb) (Choi et al. 2021; Misran et al. 2022). The value of K_F and 1/n calculated from Figure 4 was obtained as 111.45 ((mg/g) (L/mg)^1/n) and 0.28, respectively. This confirmed that PSCG had excellent ability to adsorb MB and exhibited excellent performance under various conditions. Moreover, the Temkin constant (B) was calculated as 29.45 (J/mol) and the adsorption of MB by PSCG was closer to chemisorption (B > 20 J/mol) than physical adsorption due to the change in the properties of the adsorbent by chemical substances (Table 3).

Temperature affects the adsorption equilibrium and adsorption rate as it affects the solubility of molecules and the diffusion of adsorbent materials within the pores (Choi 2022). The calculated thermodynamic parameters are represented in Table 4 and depicted in Figure 4. The negative ΔG° confirmed that the adsorption of MB onto PSCG was spontaneous and thermodynamically favorable. The positive ΔH° and ΔS° values showed that the adsorption process was endothermic in nature, and designated the affinity of PSCG and MB and increased disorder at the solid/liquid interface during the adsorption process.

The q_m of the Langmuir model is the maximum adsorbent capacity per unit mass of adsorbent, and adsorption capacity is a key parameter for evaluating the adsorption effect of an adsorbent (Leng et al. 2019; Yang et al. 2020). The maximum adsorption capacity of PSCG was 188.68 mg/g; Table 5 shows the comparison of the maximum adsorption capacity of various adsorbents based on these research results. Although the adsorption capacity of MB using PSCG did not boast a very high adsorption capacity compared to some modified lignocellulose adsorbents, it showed an adequate adsorption capacity with an appropriate amount of adsorbent. Comparison of adsorption capacity with various adsorbents showed that PSCG was a promising adsorbent for the removal of MB.

Various functional groups on the adsorbent surface play important roles in binding the adsorbent surface to the adsorbate (Gomora-Hernandez et al. 2020). There are various functional groups that can adsorb metal cations, as reported in the literature (Leng et al. 2019; Qiao et al. 2021). For example, O- or N- donors can attract hydrogen bonds and electrostatically attract harmful substances in aqueous solution due to a single pair of oxygen or nitrogen atoms (Zhuang et al. 2020; Choi et al. 2021; Elgarahy et al. 2021). Therefore, polymers containing carboxylic acids and phosphonic acids have been widely used to adsorb various cations such as heavy metals and dyes (Leng et al. 2019; Qiao et al. 2021). PSCG, a strong acid resin, adsorbs almost all cations contained in the aqueous solution by exchanging hydrogen ions due to the phosphoric acid group contained in the adsorbent. Since this ion exchange process is reversible, resins can be regenerated using strong acids when the ion exchange capacity is exhausted.

In FT-IR analysis, the amount of carboxyl (–COOH) and hydroxyl (-OH) functional groups of PSCG increased, and similar groups such as P = O and P = OOH attached to the PSCG surface interacted with MB ions in aqueous solution. Therefore, the adsorption of MB ions by the modified PSCG was achieved by the surface complexing mechanism between the oxygen-containing functional group and MB ions, which can be summarized as hydrogen bonding, electrostatic attraction, and π-π tracking interaction (Figure 5).

Analysis by FT-IR and SEM showed that phosphorylation increased the porosity and surface area of PSCG. In addition, the phosphoric acid group increased the exposure of anions on the surface by expanding the pore shape of the adsorbent, and as a result, the adsorption efficiency of MB, a cationic ion, for PSCG was improved. Elgarahy et al. (2021) reported that phosphoric acid-containing surface groups are important for the adsorption of MBs in solution. These results were consistent with the previously discussed results showing that the adsorption capacity of MB clearly correlates with the amounts of carboxylic groups and phosphate and pyrophosphate groups, respectively. The process of adsorbing MB using an adsorbent modified with phosphoric acid in aqueous solution can be summarized as follows.

• PSCG − OH + C₁₆H₁₈ClN₃S·3H₂O → PSCG − O − C₁₆H₁₈N₃S + HCl + 3H₂O

• PSCG − NH₂ + C₁₆H₁₈ClN₃S·3H₂O → PSCG − NH − C₁₆H₁₈N₃S + HCl + 3H₂O

• PSCG − COOH + C₁₆H₁₈ClN₃S·3H₂O → PSCG − COO − C₁₆H₁₈N₃S + HCl + 3H₂O

• PSCG − SO₃H + C₁₆H₁₈ClN₃S·3H₂O → PSCG − SO₃ − C₁₆H₁₈N₃S + HCl + 3H₂O

• PSCG − PO₃H + C₁₆H₁₈ClN₃S·3H₂O → PSCG − PO₃ − C₁₆H₁₈N₃S + HCl + 3H₂O

As a result of the reuse test, PSCG showed excellent performance with a high removal efficiency of 90% up to four consecutive uses. MB removal efficiency of the reused PSCG decreased to 76.5% after five consecutively times and decreased sharply to 52.4% when used the 6^th time (Figure 6). According to previous studies (Leng et al. 2019; Han et al. 2020; Choi et al. 2021), desorption was performed using HCl or NaOH for recovery of used the adsorbent. However, PSCG showed more than 90% removal efficiency up to four times without the desorption process. These results confirm that there was no doubt that PSCG was a promising adsorbent for MB adsorption. Adsorbent reusability is an important factor in determining commercial viability for industrial applications. In addition, the availability of coffee grounds, an abundant lignocellulose-based biosorbent that is readily available everywhere, is another key factor in ensuring the continuity of PSCG production. Moreover, the modification method applied to production has the advantage that it is very simple.

In this study, phosphorylated coffee grounds, a mass-produced waste was used to adsorb MB through various parameters in aqueous solution. The experimental results were analyzed by adsorption isotherms and adsorption kinetics, and the properties of adsorbents were used to explain the adsorption mechanism. After activation with phosphoric acid, PSCG had a mesopore range of 96.9%, confirming that the pore size of the adsorbent and the surface area of the adsorbent were overall increased. According to the results of EDS and pHpzc analysis, it was confirmed that acidic groups were strongly attached to the surface of PSCG. 1 g/L of PSCG adsorbent was able to remove 99% within 20 minutes, and the adsorption efficiency of MB for PSCG was 2.5–3.5 times higher than that of SCG, and it was confirmed that maximum adsorption was reached within a short time. The MB adsorption process in aqueous solution using PSCG was more suitable for the PSO and Langmuir models, and the adsorption process was closer to chemisorption than physical adsorption. The maximum MB adsorption capacity for PSCG according to the Langmuir model was 188.68 mg/g. As a result of the reuse test, PSCG showed excellent performance with a high removal efficiency of 90% up to four consecutive uses. FT-IR analysis showed that phosphorylation improved the number of carboxyl groups, and increased porosity and surface area of the adsorbent. In addition, the phosphoric acid group increased the exposure of anions on the surface by expanding the pore shape of the adsorbent, and as a result, the adsorption efficiency of MB, a cationic ion, for PSCG was improved.

This study was supported by the Basic Science Research Program through the National Research Foundation (NRF) of Korea, funded by the Ministry of Education, Science and Technology (2021R111A305924311).

The authors declare that they have no conflicts of interest.

No conflicts, informed consent, human or animal rights are applicable.

All authors whose names are listed in this manuscript certify that they have participated sufficiently in this work to take public responsibility for the content, including participation in the concept, design, data, collection, analysis, writing, and revision of the manuscript.

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