Direct functionalization of raw coal using p-xylene-2-sulfonic acid hydrate for high adsorption of bisphenol A and 4-chlorophenol: Experimental and statistical physics insights

Abstract

In this study, direct sulfonation of un-carbonized raw coal using p-xylene-2-sulfonic acid hydrate was successfully achieved—a rarely reported modification route compared with the conventional sulfonation of activated carbon after carbonization. This one-step molecular-level substitution presents a key novelty, enabling transformation of the native turbostratic microstructure without prior thermal treatment. XRD analysis confirmed substantial disruption of semi-crystalline carbon domains, with strong suppression of the characteristic (10) band at 40–50° (2θ), indicating conversion to a highly amorphous and defect-rich surface. The resulting sulfonated coal (SF.CA) demonstrated markedly enhanced adsorption performance toward Bisphenol A and 4-chlorophenol, reaching Q_sat values of 245.78 mg/g and 242.44 mg/g, respectively, positioning the material among the most effective carbonaceous sorbents for phenolic pollutants. Equilibrium modeling identified the single-site monolayer model as the best fit (R² = 0.999; χ² as low as 0.03). Statistical-physics analysis provided deeper mechanistic insight, revealing multi-molecular stacking at each adsorption site, with steric coefficients n = 3.21–3.97 and active-site densities Nₘ decreasing from 76.75 to 47.90 mg/g with increasing temperature. This decline directly reflects temperature-sensitive anchoring behavior. The negative energetic parameter (ΔE down to −15.06 kJ/mol) across all temperatures confirms a spontaneous, low-energy, exothermic adsorption mechanism. The combined steric and energetic parameters support a vertical multi-molecule assembly stabilized through hydrogen bonding and π–π interactions at the SF.CA interface. Overall, the findings show that direct sulfonation is both synthetically simpler and capable of producing a highly efficient adsorbent whose behaviour is validated quantitatively at the molecular scale.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-63296-3.

Introduction

Water availability and human security stand among the most pressing global challenges of the twenty-first century^1,2. The World Health Organization (WHO) has issued an urgent warning that, by 2025, nearly half of the world’s population will experience conditions of water scarcity^2–4. The unprecedented pace of industrialization over recent decades has intensified environmental pressures, particularly through water contamination, which threatens both public health and aquatic ecosystems^5,6. A major concern in this context is the presence of endocrine-disrupting chemicals (EDCs)—exogenous compounds capable of interfering with hormonal pathways essential for growth, development, and reproduction in living organisms^7,8. Exposure to EDCs has been associated with a wide range of disorders and congenital abnormalities. Notably, the European Union has classified more than 900 substances, including numerous phenolic derivatives, as potential EDCs^7–9.

Phenolic organic compounds are among the most hazardous classes of waterborne pollutants, largely introduced as secondary products from the synthesis of insecticides, petroleum-based materials, and wood-processing industries¹⁰. A notable example is bisphenol A (BSP-A) [2,2-bis(4-hydroxyphenyl) propane], a widely manufactured phenolic derivative categorized as an endocrine-disrupting chemical (EDC)^10,11. BSP-A has broad industrial applications, serving as a key precursor in the production of epoxy resins, synthetic fibers, electronics, healthcare devices, and leather-tanning agents. However, its widespread use has led to its detection as a micropollutant in aquatic environments, where it has been linked to severe health concerns, including malignancies, developmental defects, endocrine disorders, and elevated risks of metabolic and cardiovascular diseases such as diabetes, obesity, and heart conditions^10,12. To mitigate these risks, international regulatory agencies advise limiting BSP-A concentrations in human tissues to below 50 μg/kg¹³, while water quality guidelines in regions such as the United States and China restrict maximum allowable levels to 10 μg/L^9,14.

Chlorophenols (CLs) represent another critical group of phenolic contaminants, widely applied in pesticide, fungicide, and chemical manufacturing^15,16. Their persistence, strong bioaccumulation potential, and pronounced toxicity make them particularly dangerous to aquatic ecosystems and public health, even at trace concentrations^17,18. Extensive research has demonstrated that CLs exhibit mutagenic, carcinogenic, and genotoxic properties, coupled with resistance to natural biodegradation, thereby contributing to their long-term ecological and toxicological impact^15,16,19. Among these compounds, 4-chlorophenol (4-CL) is heavily employed across medicinal, petrochemical, pigment, and industrial chemical sectors²⁰. Due to its toxicity, stability, and carcinogenic potential, the U.S. Environmental Protection Agency has listed 4-CL as one of the most hazardous environmental contaminants¹⁸. Consequently, developing advanced and sustainable technologies for the effective elimination of both BSP-A and 4-CL from contaminated water sources remains a critical global priority.

The remediation of hazardous pollutants such as 4-chlorophenol (4-CL) and bisphenol A (BSP-A) has been pursued through multiple approaches, including photocatalytic degradation, biological treatments, adsorption processes, and advanced oxidation technologies^21,22. Among these, adsorption has emerged as one of the most widely investigated and practically applied methods, owing to its cost-effectiveness, operational simplicity, high efficiency, reusability, and adaptability for large-scale applications^23,24. The effectiveness of an adsorbent, however, is determined by a range of critical parameters, including synthesis cost, ease of fabrication, raw material availability, adsorption capacity, regeneration potential, kinetic performance, long-term sustainability, selectivity, safety, and chemical stability^25,26. In response to these demands, extensive research efforts have been devoted to the design and optimization of advanced adsorbents derived from inexpensive and abundantly available feed stocks, particularly natural minerals and agricultural or industrial byproducts^27,28. The adoption of such environmentally benign and resource-efficient precursors is strongly encouraged, as they offer significant ecological and economic benefits while ensuring the sustainable application of adsorption technologies²⁹.

In recent years, different ranks of natural coal have been extensively examined as potential adsorbent materials with remarkable capabilities for removing dissolved pollutants from water, particularly synthetic dyes^30–32. In addition to dyes, coal and coal-derived products have been successfully applied for the elimination of pesticides and pharmaceutical residues^33,34, confirming their broad applicability toward diverse classes of organic contaminants. Structurally, coal is a complex sedimentary rock composed predominantly of organic macromolecular components such as cellulose, lignin, resins, and macerals, along with varying proportions of inorganic matter, including mineral impurities and ash^32,35,36. From a chemical standpoint, coal is largely constituted of polycyclic aromatic hydrocarbons arranged in disordered networks of fused aromatic rings that are bridged by aliphatic chains and linked with oxygenated functional groups, including hydroxyl, carbonyl, phenolic, and carboxyl moieties^37,38. These oxygen-containing groups play a pivotal role in surface reactivity, providing active sites for interactions with pollutants through different types of interaction mechanisms³⁹. The heterogeneous distribution of such functional groups, coupled with the porous and layered nature of coal, enhances its affinity toward a wide spectrum of dissolved organics. Moreover, the variation in coal rank directly influences its physicochemical properties—higher-rank coals typically exhibit more condensed aromatic structures and lower oxygen content, whereas lower-rank coals retain greater numbers of oxygenated functionalities and higher porosity, making them particularly effective in adsorption applications^32,36,38. Taken together, these structural and chemical attributes highlight coal as a multifunctional, naturally abundant material with significant potential for use as a low-cost and sustainable adsorbent in water purification technologies.

The physicochemical and adsorption behavior of coal is largely governed by its surface chemistry and textural features, particularly the abundance and distribution of active functional groups^32,40. To improve these properties and expand its application as an efficient adsorbent, numerous functionalization and modification strategies have been explored. Such methods aim to optimize surface reactivity, increase porosity, and enhance pollutant-binding capacity. Common approaches include chemical activation, demineralization, thermal treatment, and surface loading with metals or metal oxides^31,32. Among these, chemical activation has been extensively employed to improve the performance of carbonaceous materials by promoting the generation or exposure of reactive surface groups. Oxygenated functionalities—primarily carboxyl (–COOH) and hydroxyl (–OH) moieties—play a central role in pollutant binding, as they provide active adsorption sites capable of interacting with dissolved contaminants through different mechanisms and complexation^30,41.

Sulfonation has emerged as a promising strategy for improving the surface reactivity of coal-based adsorbents. The introduction of sulfonic acid groups (–SO₃H) increases surface acidity and electronegativity, promotes oxidation of native functional groups, and generates oxygenated active sites, thereby enhancing adsorption affinity toward contaminants^30,41. However, sulfonation efficiency depends strongly on the sulfonating reagent, temperature, acid concentration, reaction time, and homogenization conditions^38,42. These factors determine the density and stability of introduced acidic groups, as well as the extent of structural disruption, demineralization, and exposure of adsorption-active domains in the coal matrix. Previous coal sulfonation studies have confirmed the beneficial role of –SO₃H enrichment, but they have mainly relied on concentrated H₂SO₄ as the sulfonating/oxidizing agent and have commonly focused on catalysis or non-phenolic contaminants. For instance, coal-derived solid acid catalysts have been prepared through carbonization or partial carbonization followed by H₂SO₄ sulfonation and applied mainly in esterification or biodiesel-related reactions^35,36. Other studies treated raw or sub-bituminous coal with H₂SO₄ to produce acidic catalysts, where the emphasis was acid density and transesterification activity rather than adsorption of phenolic pollutants⁴⁰. In adsorption applications, H₂SO₄-treated low-rank coal has been reported for methylene blue removal³⁰, H₂SO₄-sulfonated natural coal for malachite green adsorption⁴³, and sulfonated raw coal for methyl parathion adsorption⁴⁴. Thus, previous routes remain dominated by H₂SO₄-based modification, carbonization-assisted sulfonation, catalytic applications, dye removal, or pesticide adsorption.

In contrast, the present study introduces direct sulfonic-acid functionalization of raw, uncarbonized sub-bituminous coal using analytical-grade p-xylene-2-sulfonic acid hydrate (C₈H₁₀O₃S·xH₂O) as an alternative sulfonating reagent. Unlike conventional H₂SO₄-based approaches, this treatment is performed directly on the native coal framework without prior carbonization or activation. Therefore, the novelty of this work is not merely the preparation of sulfonated coal, but the direct modification of uncarbonized coal through a sulfonic-acid route and its evaluation for adsorption of two toxic phenolic pollutants, bisphenol A (BSP-A ) and 4-chlorophenol (4-CL). This approach is expected to induce simultaneous –SO₃H incorporation, oxygenated-group enrichment, partial demineralization, and disruption of the native turbostratic coal structure, thereby increasing the number and accessibility of active adsorption sites. The adsorption performance of the resulting sulfonated coal is evaluated using classical kinetic/isotherm models and statistical physics modeling to quantify steric and energetic parameters. Coupling direct functionalization with molecular-scale adsorption analysis provides mechanistic insight into phenolic pollutant–coal interactions and establishes p-xylene-2-sulfonic acid-modified raw coal as a promising low-cost adsorbent for water purification.

Methodology

Materials

The natural coal employed in this study was sub-bituminous coal obtained from the El Maghara coal mine located in Sinai, Egypt. Its chemical composition was characterized through both proximate and ultimate analyses, with the detailed results provided in Table S1. For surface modification, analytical-grade p-xylene-2-sulfonic acid hydrate (PXSA; 2,5-dimethylbenzenesulfonic acid hydrate; IUPAC: 2,5-dimethylbenzene-1-sulfonic acid; C₈H₁₀O₃S·xH₂O; CAS No. 609-54-1; MW: 186.23 g/mol, anhydrous basis; Sigma-Aldrich) was used as the organic sulfonic-acid modifier. The xH₂O term denotes variable hydration. PXSA was selected for its strongly acidic –SO₃H group, aromatic xylyl compatibility with coal domains, and capacity to increase Brønsted acidity and oxygenated functionalities through direct modification of raw coal without prior carbonization or activation, instead of conventional H₂SO₄ treatment, for pollutant adsorption. Bisphenol A (BSP-A , >99% purity; Sigma-Aldrich) and 4-chlorophenol (4-CL, 98% purity) were selected as model phenolic pollutants for the preparation of contaminated aqueous solutions. All chemicals utilized in this work were of analytical grade and were applied without any further purification.

Sulfonation modification

Coal sulfonation was conducted according to the procedure of Ibrahim et al.²⁴ with minor modifications. The supplied coal samples were first ground to a particle size of 20–70 µm using a laboratory grinder to improve surface exposure. A portion of 10 g of the ground coal was then mixed with 100 mL of molten p-xylene-2-sulfonic acid hydrate (PXSA) in a 250 mL flask. The mixture was maintained at 90 °C for 48 h under continuous magnetic stirring (500 rpm), while ultrasonic irradiation was applied at fixed intervals to promote homogenization and enhance contact between the acid and coal particles. After treatment, the slurry was cooled to room temperature and repeatedly washed with distilled water until neutral pH was achieved, ensuring the removal of excess sulfonic acid and preventing surface over-acidification. The washed samples were then dried in a digital oven at 60 °C for about 10 h to obtain stable powders. The resulting sulfonated coal materials were stored in sealed tubes under dry conditions and labelled as SF.CA for further characterization and adsorption experiments (Fig. 1).

Fig. 1.

Schematic illustration of raw coal functionalization using p-xylene-2-sulfonic acid hydrate (PXSA) and preparation of sulfonated coal (SF.CA).

Analytical techniques

To evaluate the influence of sulfonation, the modified and unmodified coal samples were subjected to a series of complementary characterization methods targeting their structural, chemical, morphological, and textural attributes. The crystal structure was examined by recording X-ray diffraction (XRD) patterns with a PANalytical Empyrean diffractometer, which allowed the identification of crystalline components and the detection of structural rearrangements introduced during treatment. The surface functional chemistry was probed using Fourier transform infrared spectroscopy (FTIR; Shimadzu FTIR-8400S). This technique was applied to track changes in the native functional groups of coal and to confirm the incorporation of sulfur-bearing moieties, with emphasis on diagnostic vibrational bands of S=O, C=O, O–H, and C–O. The quantitative distribution of acidic surface groups was determined through Boehm titration following standard procedures, providing direct measurements of carboxylic, phenolic, and p-xylene-2-sulfonic acid functionalities generated or enhanced during modification. Alterations in surface morphology and microstructure were visualized by scanning electron microscopy (SEM; Zeiss Gemini Ultra 55), which revealed modifications in particle texture, surface roughness, and porosity development as a consequence of sulfonation. In addition, the textural properties were assessed by nitrogen adsorption–desorption isotherms using a Beckman Coulter SA3100 surface area analyzer. The Brunauer–Emmett–Teller (BET) surface area, total pore volume, and pore size distribution were calculated to quantify the impact of sulfonation on accessible porosity and surface area.

Adsorption studies

The adsorption performance of raw coal (CA) and sulfonated coal (SF.CA) toward bisphenol A (BSP-A ) and 4-chlorophenol (4-CL) was examined through a series of batch experiments under systematically varied conditions. The parameters investigated included solution pH (3–10), contact time (20–1080 min), initial pollutant concentration (50–400 mg/L), and temperature (303–323 K). For consistency, three factors were held constant throughout all trials: an adsorbent dosage of 0.25 g/L, a solution volume of 100 mL, and a fixed agitation speed. Each experimental run was performed in triplicate, and the mean values were reported to ensure reproducibility and statistical reliability.

At the completion of each equilibration period, the adsorbent particles were separated from the solutions treated by filtration using Whatman filter paper. The residual concentrations of BSP-A and 4-CL were quantified by high-performance liquid chromatography (HPLC; Merck/Hitachi), equipped with a Phenomenex Luna PFP column (150 mm × 4.6 mm, 5 µm particle size). The system comprised an L-7400 UV detector, an L-7100 pump, a Rheodyne 7725i injection valve, and a 20 µL sampling loop, ensuring accurate detection of the target phenolic compounds.

The equilibrium adsorption capacity (Qe, mg/g) was determined according to Equation (1):

1

where C₀ and (mg/L) are the initial and equilibrium pollutant concentrations, V (mL) is the solution volume, and m (g) is the adsorbent mass. This relationship allowed for a quantitative evaluation of the removal efficiency of both CA and SF.CA under the studied conditions.

The adsorption kinetics and equilibrium behavior were analyzed by fitting the experimental data to conventional kinetic and isotherm models using non-linear regression (Table S1). The adequacy of these models was judged by the determination coefficient (R²) (Eq. 2) and the Chi-squared error function (χ²) (Eq. 3).

2

3

To gain deeper insights into the adsorption mechanisms, advanced isotherm modeling was carried out within the framework of statistical physics theory. This approach enabled the evaluation of steric and energetic parameters by non-linear fitting of experimental data to advanced equations (Table S1). The quality of these fittings was judged using the correlation coefficient (R²) and the root mean square error (RMSE) (Eq. 4):

4

where ,cal and ,exp represent the calculated and experimental adsorption capacities, respectively, denotes the number of experimental data points, and p corresponds to the number of model parameters. This advanced modeling framework provided a comprehensive description of the adsorption process and a robust assessment of the predictive performance of the applied models.

Results and discussion

Characterisation of the adsorbent

XRD analysis

The powder X-ray diffraction (XRD) patterns of the raw coal and the sulfonated coal (SF.CA) demonstrate a clear structural transformation from a semi-crystalline turbostratic carbon framework to a predominantly amorphous configuration following sulfonation (Fig. 2). In the raw coal (Fig. 2A), two broad diffraction features appear within the ranges of approximately 8–30° and 35–50° (2θ, Cu Kα). The low-angle region corresponds to the (002) reflection of carbon, which reflects the out-of-plane stacking of aromatic layers, whereas the higher-angle region represents the so-called “10 band,” comprising overlapping (100)/(101) in-plane reflections characteristic of disordered graphitic (turbostratic) carbons. The broadness and low intensity of these diffraction bands indicate limited layer stacking and short lateral coherence lengths of small graphitic microdomains embedded within a largely amorphous matrix—an XRD pattern commonly observed for coals and activated carbons^45–47.

Fig. 2.

XRD pattern of starting coal sample (A). and sulfonated product using p-xylene-2-sulfonic acid (SF.CA) (B).

After sulfonation, the XRD profile undergoes a pronounced transformation. The diffraction pattern becomes dominated by a broad, diffuse halo extending from approximately 15° to 30° (2θ, Cu Kα), while the (10) band in the 40–50° range is significantly diminished (Fig. 2B). This broadening and intensity loss are characteristic indicators of heightened structural disorder in carbon materials, signifying reduced stacking coherence (smaller Lc values) and disrupted lateral order, rather than effects arising from preferred orientation^47–49.These structural changes align with the chemical modifications introduced during sulfonation, where –SO₃H and other oxygen-containing functional groups are incorporated at defect sites and edges of the carbon framework. Such substitutions interfere with interlayer registry and weaken the π–π interactions among aromatic layers. Comparable disordering effects and the resulting “two-band” diffraction signature have been widely documented in acid-treated and sulfonated carbons^48,50,51. Additionally, a weak and broad reflection around 2θ ≈ 63°, absent in the raw sample, is likely attributable to trace mineral impurities—such as α-quartz, which exhibits reflections in the 60–65° region—rather than higher-order graphitic peaks (noting that the (004) reflection of graphite appears near ~55° 2θ with Cu Kα). This minor high-angle feature further underscores the predominance of amorphous domains in the SF.CA sample while suggesting that small quantities of residual crystalline impurities or nanoscopic graphitic regions may still persist within the amorphous matrix.Taken together, the dominant amorphous halo and suppression of the (10) band in SF.CA provide strong evidence that sulfonation disrupts the semi-crystalline turbostratic framework, yielding a highly amorphous, defect-rich, and functionalized carbon. This increased disorder enhances the density of defect sites and exposes polar functional groups, properties that are strongly associated with improved adsorption of phenolic and other polar solutes.

FT-IR analysis

Figure 3 shows the FT-IR spectra of raw coal (A) and sulfonated coal (B, SF.CA). The raw coal spectrum exhibits several hydroxyl-related absorptions at 3755, 3655, 3551, 3486, 3415, 3234, and 3117 cm⁻¹, which can be assigned to free or structural –OH groups of clay minerals, hydrogen-bonded hydroxyl groups, phenolic –OH, and adsorbed water; minor N–H stretching contributions may also overlap in the 3300–3500 cm⁻¹ region^52–54. The bands at 2923 and 2850 cm⁻¹ correspond to asymmetric and symmetric aliphatic C–H stretching vibrations of –CH₂/–CH₃ groups, confirming the presence of aliphatic moieties in the coal macromolecular structure^47,55,56. The absorptions at 1640 and 1617 cm⁻¹ are attributed to H–O–H bending of adsorbed water and aromatic C=C skeletal vibrations. The bands at 1460 and 1401 cm⁻¹ are associated with aliphatic C–H deformation and aromatic ring vibrations. In the fingerprint region, the absorptions at 1166 and 1001 cm⁻¹, together with bands at 816, 625, 603, 532, 474, and 425 cm⁻¹, are mainly related to Si–O–Si and Si–O–Al vibrations of quartz and clay minerals, especially kaolinite, with possible overlap from C–O stretching of phenolic, ether, and alcohol groups^57–59. The weak band around 2370 cm⁻¹ is commonly associated with atmospheric CO₂ and is therefore not considered a diagnostic coal structural band.

Fig. 3.

FT-IR spectra of starting coal sample (A) and sulfonated product using p-xylene-2-sulfonic acid (SF.CA) (B).

After sulfonation, the spectrum of SF.CA shows clear spectral changes. The broad O–H stretching envelope becomes more intense and shifts to 3399 cm⁻¹, indicating stronger hydrogen bonding due to the introduction of acidic –SO₃H groups, additional oxygenated surface sites, and enhanced water affinity^47,50,60. The residual band at 3755 cm⁻¹ may arise from remaining free or structural hydroxyl groups. The broad absorptions at 2494 and 2381 cm⁻¹ may reflect strongly hydrogen-bonded acidic O–H species, although the latter can overlap with atmospheric CO₂. The aliphatic C–H bands at 2923 and 2850 cm⁻¹, clearly present in raw coal, are strongly reduced or nearly absent fafter treatment, suggesting partial oxidation, cleavage, or removal of aliphatic fragments and relative enrichment of aromatic/oxygenated domains^61,62.

The most important evidence for sulfonation appears in the fingerprint region of SF.CA. New or intensified bands at 1289 and 1175 cm⁻¹ are assigned to asymmetric O=S=O stretching vibrations of aryl-sulfonic acid/sulfonate groups, while bands at 1068 and 1011 cm⁻¹ correspond to symmetric O=S=O and S–O stretching vibrations, with possible overlap from C–O and residual Si–O modes^40,63–65. The band at 880 cm⁻¹ is attributed to S–O–H deformation, whereas the absorptions at 659 and 579 cm⁻¹ are assigned to C–S and/or S–O vibrations characteristic of aryl-sulfonated structures^40,63–65. The remaining low-wavenumber band at 455 cm⁻¹ indicates residual mineral-related Si–O/Al–O bending. Compared with raw coal, the mineral-associated bands become less distinct and partially overlapped by sulfonic-acid vibrations, supporting partial acid leaching and modification of the mineral fraction^53,66.

It should be noted that a well-resolved carbonyl band at 1715–1690 cm⁻¹ is not clearly observed in the corrected spectra; therefore, formation of carboxylic C=O groups should be discussed cautiously. Overall, the reduction of aliphatic C–H bands, intensification of the broad hydroxyl region, appearance of diagnostic O=S=O, S–O–H, and C–S/S–O absorptions, and attenuation of mineral-related signals collectively support successful surface sulfonation of coal and partial acid-induced modification of the coal matrix. These changes are expected to increase surface acidity, polarity, and hydrogen-bonding capacity, thereby enhancing interactions with phenolic adsorbates.

Energy-Dispersive X-ray spectroscopy (EDX)

Figure S1 displays the EDX spectra of raw coal (A) and the sulfonated product (B). Raw coal is mainly composed of carbon and oxygen, with average contents of 66.77 ± 0.67 wt.% and 25.43 ± 0.95 wt.%, respectively. Minor inorganic constituents are also detected, including Si, Fe, and native sulfur, with average contents of 0.73 ± 0.06, 2.97 ± 0.29, and 4.13 ± 0.25 wt.%, respectively.

After sulfonation, the elemental composition changes markedly. The sulfur content increases strongly from 4.13 ± 0.25 wt.% in raw coal to 24.32 ± 6.28 wt.% in SF.CA, corresponding to an increase from 1.77 ± 0.12 to 12.52 ± 3.81 at.%. This substantial sulfur enrichment confirms the successful incorporation of sulfur-bearing functional groups, especially –SO₃H groups, onto the coal surface. In parallel, the oxygen content increases from 25.43 ± 0.95 to 40.02 ± 3.88 wt.%, supporting the formation or enrichment of oxygenated functionalities such as –SO₃H, –OH, –COOH, and C–O groups. The carbon content decreases from 66.77 ± 0.67 to 34.20 ± 5.48 wt.%, which is consistent with the increased contribution of oxygen- and sulfur-rich surface groups after modification.

The inorganic elements also decrease after treatment, particularly Si from 0.73 ± 0.06 to 0.18 ± 0.30 wt.% and Fe from 2.97 ± 0.29 to 0.68 ± 0.47 wt.%, indicating partial acid-assisted demineralization and depletion of mineral impurities from the coal surface. Trace Al, Ca, and F are detected only in the modified sample at very low levels, reflecting the heterogeneous nature of the residual mineral fraction. Overall, the quantitative EDX results confirm that sulfonation caused strong sulfur enrichment, oxygenation of the coal surface, and partial removal of inorganic ash components.

Surface morphology (SEM)

Figure 4 illustrates the morphological contrast between raw coal (A–C)^32,52 and the sulfonated product (D–F). The raw coal exhibits compact, lamellar packets formed by stacked plate-like flakes. At the 10 μm scale, the aggregates appear dense and relatively smooth, with limited edge roughness and few interparticle voids (Fig. 5A-C)^32,52. Higher magnification reveals broad, flat facets separated by step-like ledges and occasional cleavage cracks, with scattered bright specks along edges and gaps that correspond to mineral inclusions acting as cement between lamellae (Fig. 4A-C)^32,52. This tightly packed flake morphology reflects a low abundance of external roughness and a restricted number of accessible surface defects. In contrast, sulfonation induces profound surface reworking (Fig. 4D-F). The lamellae disintegrate into friable, irregular fragments with jagged edges and pervasive microcracking. Smooth plate-like surfaces are replaced by etched and pitted textures, developing a granular, cauliflower-like crust composed of submicron debris and loosely bound aggregates (Fig. 4D-F). Numerous nano- to submicron cavities appear both within and between particles, and the matrix becomes deflocculated with widened intergranular corridors.

Fig. 4.

SEM micrographs of (A–C) raw coal^32,52 and (D–F) coal sulfonated with p-xylene-2-sulfonic acid.

These features are consistent with the chemistry of p-xylene-2-sulfonic acid treatment: protonation and electrophilic sulfonation at defect sites weaken π–π stacking and promote delamination, while acid-catalyzed cleavage of aliphatic bridges and dehydration generate pits and cracks. Simultaneously, dissolution of ash components (e.g., carbonates, Fe-, and Si-bearing phases) removes cementing material, further opening voids. Bright nanometric spots in the sulfonated product correspond to residual aluminosilicate inclusions or sulfur-rich domains, consistent with the EDX findings. Overall, the SEM images demonstrate a transformation from compact, flake-stacked lamellae to a rough, porous, and fragmented microtexture enriched in edge sites and open pathways.

N₂ adsorption–desorption analysis

The textural properties of the natural coal treated with p-xylene-2-sulfonic acid hydrate were evaluated by N₂ adsorption–desorption analysis at 77 K (Fig. 5). The PXSA-treated coal exhibited a BET specific surface area of 23.6 m² g⁻¹, indicating moderate accessible surface area for chemically treated natural coal. The BET method provides an estimate of probe-accessible surface area based on multilayer gas adsorption, and for heterogeneous coal matrices this value should be considered an apparent surface area affected by pore accessibility, mineral matter, fractures, and organic maceral structure⁶⁷.

The adsorption–desorption profile can be classified as a mixed Type II/IV(a)-like isotherm with an H3-type hysteresis loop tendency. The limited N₂ uptake at low relative pressure indicates the absence of a dominant N₂-accessible microporous network, while the gradual uptake at intermediate pressure and sharp increase near p/po ≈ 0.1 suggest multilayer adsorption and capillary condensation in larger pores, cracks, and interparticle voids. According to IUPAC classification, Type II behavior is generally associated with nonporous or macroporous solids, whereas Type IV(a) is related to capillary condensation in mesopores⁶⁷. The H3-type hysteresis loop is commonly attributed to slit-shaped pores, particle aggregates, and non-rigid pore networks that do not show a clear limiting adsorption plateau at high relative pressure⁶⁷.

The pore-size distribution shows a main positive maximum at approximately 59.5 nm, indicating that the accessible pore volume is mainly located at the large-mesopore/macropore boundary. This suggests that PXSA treatment modified pore accessibility and exposed larger voids or cracks rather than producing a highly microporous structure. Negative values observed in part of the differential pore-volume curve are physically meaningless and should be treated as BJH calculation artifacts, most likely arising from weak desorption signals, baseline correction, or pore-network effects⁶⁸.

Adsorption results

Effect of pH

The pH of an aqueous system is a decisive parameter governing both the surface charge characteristics of SF.CA and the ionization state of dissolved phenolic pollutants⁶⁹. Adsorption experiments were conducted over a pH range of 3 to 10 under fixed conditions: contact time = 120 min, temperature = 303 K, solution volume = 100 mL, pollutant concentration = 100 mg/L, and adsorbent dosage = 0.25 g/L. As the pH increased from 3 to 8, the uptake capacity of bisphenol A (BSP-A) rose markedly from 15.7 to 70.5 mg/g (Fig. 6). Beyond pH 8, a distinct reduction in adsorption efficiency was observed. These findings indicate that SF.CA performs most effectively for BSP-A removal within the slightly acidic to weakly alkaline range, which lies within the pH limits recommended for industrial effluent discharge by the U.S. EPA⁷⁰.

Fig. 5.

N₂ adsorption–desorption isotherm of p-xylene-2-sulfonic acid-treated natural coal with inset pore-size distribution.

This pH-responsive behavior reflects the acid–base speciation of BSP-A and the surface chemistry of SF.CA. BSP-A exhibits two phenolic dissociation constants, pKₐ₁ = 9.6 and pKₐ₂ = 10.2⁷¹. Therefore, under moderately acidic to weakly basic conditions, particularly between pH 5 and 8, BSP-A exists predominantly in its neutral molecular form. In this range, adsorption is mainly promoted by hydrogen bonding between the phenolic –OH groups of BSP-A and the –SO₃H, –COOH, –OH, and C=O groups of SF.CA, together with π–π interactions between the aromatic rings of BSP-A and the carbonaceous domains of the modified coal^71,72. The maximum uptake near pH 8 can therefore be attributed to favorable non-covalent interactions, improved accessibility of active surface groups, and limited electrostatic repulsion, rather than direct electrostatic attraction of neutral BSP-A molecules. At pH > 9, BSP-A progressively deprotonates to form negatively charged bisphenolate species, while the SF.CA surface becomes increasingly negatively charged due to deprotonation of acidic surface groups. This generates electrostatic repulsion and weakens hydrogen-bonding contributions, resulting in the observed decrease in adsorption capacity⁷².

Similarly, pH exerts a pronounced influence on the uptake of 4-chlorophenol (4-CL). The adsorption capacity of SF.CA increased steadily up to pH 7, attaining 45.6 mg/g (Fig. 6), but declined sharply at higher pH values. In acidic media, excess H⁺ ions can compete with pollutant molecules for active sites and may reduce the availability of functional groups involved in hydrogen bonding, thereby suppressing 4-CL adsorption⁷³. 4-CL has a single phenolic dissociation constant of approximately pKₐ = 9.4⁷⁴. Thus, under acidic and near-neutral conditions, 4-CL remains mainly in its neutral molecular form, which favors adsorption through hydrogen bonding, π–π interactions, and hydrophobic/dispersion interactions with the functionalized coal surface. However, when the pH approaches and exceeds the pKₐ of 4-CL, the molecule progressively ionizes to form negatively charged 4-chlorophenolate species. These anionic species experience electrostatic repulsion from the negatively charged SF.CA surface, leading to the marked decline in adsorption under alkaline conditions⁷⁵.

Contact time

The influence of contact time on the adsorption efficiency of SF.CA toward bisphenol A (BSP-A) and 4-chlorophenol (4-CL) was systematically examined within a time interval of 20 to 1080 minutes. Throughout the experiments, all other conditions were kept constant: initial pollutant concentration of 100 mg/L, solution volume of 100 mL, temperature of 303 K, pH values of 8 (for BSP-A) and 7 (for 4-CL), and an adsorbent dosage of 0.25 g/L. As illustrated in Fig. 7A, the adsorption capacity of SF.CA for both contaminants increased rapidly with contact time, reaching a pronounced rise during the first 300 minutes. Thereafter, the adsorption rate gradually leveled off, indicating that equilibrium had been achieved. Beyond this point, only marginal changes in pollutant removal were detected, suggesting saturation of the active binding sites on the SF.CA surface.

Fig. 6.

Effect of initial solution pH on the adsorption of bisphenol A (BSP-A) and 4-chlorophenol (4-CL) by SF.CA.

The equilibrium time of SF.CA should be interpreted in relation to both adsorption capacity and mass-transfer behavior. For phenolic pollutants, reported equilibrium/contact times vary widely depending on the adsorbent origin, particle morphology, pore accessibility, and density of surface functional groups. Zhao et al.⁷⁶ reported that H₂O₂-modified wetland-plant biochars reached BSP-A adsorption equilibrium within 300 min, while Din et al.⁷⁷ observed that BSP-A adsorption onto raw and treated banana-peel adsorbents became nearly constant after approximately 200 and 150 min, respectively. In contrast, Mourão et al.⁷⁸ reported much longer BSP-A equilibrium times for activated-carbon systems, reaching 10–12 h for powdered activated carbon and approximately 24 h for granulated activated carbon, mainly because microporous constraints limited BSP-A diffusion toward internal active sites. For 4-chlorophenol adsorption, Tazik et al.⁷⁹ showed faster uptake using amine-functionalized activated carbon, where removal became almost constant after 45 min and the maximum adsorption capacity reached 316.1 mg/g. Therefore, the equilibrium time of SF.CA, approximately 300 min, falls within the reported kinetic range for phenolic-pollutant adsorption on carbonaceous and bio-based adsorbents.

At equilibrium, the maximum uptake capacities of SF.CA were calculated as 120.8 mg/g for BSP-A and 92.4 mg/g for 4-CL, confirming its high adsorption potential under the conditions studied (Fig. 7A). The initial stage of the process was characterized by a steep increase in removal efficiency, attributable to the abundant availability of reactive and unoccupied sites on the SF.CA surface that allowed rapid attachment of pollutant molecules⁸⁰. As adsorption progressed, these active centers became progressively occupied, reducing the number of accessible sites and thereby slowing the adsorption rate. Eventually, equilibrium was reached when all functional sites were saturated and further increases in contact time did not yield noticeable enhancement in pollutant uptake⁸¹. This time-dependent behavior reflects a typical two-phase adsorption pattern: a fast initial diffusion and surface binding phase governed by abundant available sites, followed by a slower equilibrium phase controlled by the decreasing number of vacant active centers. Such kinetics confirms that the adsorption of both BSP-A and 4-CL on SF.CA is primarily surface-controlled and reaches equilibrium within approximately five hours under the present experimental conditions.

Kinetic studies

Intra-particle diffusion behavior

The intra-particle diffusion analysis provides crucial insights into the adsorption pathways and molecular interactions governing the uptake of bisphenol A (BSP-A) and 4-chlorophenol (4-CL) by SF.CA. As illustrated in Fig. 7B, the adsorption profiles exhibit three distinct linear regions, each corresponding to a specific adsorption stage characterized by different diffusion rates. These segments collectively describe the sequential nature of the adsorption process, involving external surface adsorption, internal diffusion, and final equilibrium attainment⁸². The first phase—external adsorption—is dominated by rapid binding of BSP-A and 4-CL molecules onto the readily accessible active sites on the outer surface of SF.CA particles. This initial step proceeds at a high rate owing to the large number of vacant surface sites and strong electrostatic and π–π interactions between the pollutant molecules and the functional groups of the adsorbent. Consequently, this stage controls the early kinetics and contributes most significantly to the overall removal efficiency⁸³.

Following the initial phase, adsorption transitions into the internal diffusion stage, where pollutant molecules progressively migrate from the external surface into the inner pore structure of SF.CA (Fig. 7B). During this stage, molecular diffusion becomes the rate-limiting step as BSP-A and 4-CL molecules penetrate deeper into the adsorbent matrix and occupy interior adsorption sites. The gradual decline in the adsorption rate observed in this phase reflects the slower diffusion of molecules through internal pore channels, coupled with increased resistance to mass transfer within the adsorbent framework^83,84. Finally, the process enters the equilibrium or saturation stage, where the adsorption capacity reaches a steady state. At this point, nearly all accessible binding sites—both external and internal—are occupied, and the rate of adsorption approaches zero. Any residual uptake beyond this phase is attributed to weak van der Waals interactions and intermolecular attractions rather than active site binding^85,86.

Overall, the observed multi-step diffusion behavior reveals that adsorption of BSP-A and 4-CL on SF.CA proceeds through a combination of surface adsorption and intra-particle diffusion mechanisms. The dominance of the initial external phase followed by slower internal diffusion underscores the dual contribution of surface reactivity and porous structure to the overall adsorption efficiency. This well-defined progression highlights the effectiveness of SF.CA in utilizing both its external surface and internal porosity for pollutant sequestration, thereby achieving high removal performance

Kinetic modeling

A comprehensive understanding of the adsorption kinetics of bisphenol A (BSP-A) and 4-chlorophenol (4-CL) onto SF.CA is crucial for identifying the rate-determining mechanisms and distinguishing between the physical and chemical interactions that govern the overall process. The adsorption rate is influenced by several concurrent phenomena, including external mass transfer, intra-particle diffusion, and surface interactions, along with potential chemical processes such as hydrogen bonding or complexation between the pollutant molecules and functional groups on the adsorbent surface⁸⁷. To evaluate these kinetic characteristics, two classical models, the pseudo-first order (P.F.) and pseudo-second order (P.S.) models—were applied to experimental data. The pseudo-first-order model, originally proposed by Lagergren, assumes that the rate of adsorption is directly proportional to the number of available adsorption sites on the adsorbent surface⁸⁰. It typically describes systems where physisorption predominates, driven mainly by weak van der Waals forces and electrostatic attractions between the adsorbate and the adsorbent⁸⁸. In contrast, the pseudo-second-order model, developed by Ho and McKay, assumes that adsorption proceeds primarily through chemisorption involving valence electron sharing or exchange between the active sites of the adsorbent and the adsorbed molecules⁸⁹.

Non-linear regression analyses were performed to fit the experimental data to both kinetic models, and the quality of each fit was assessed using the determination coefficient (R²) and chi-squared (χ²) parameters as indicators of accuracy. As shown in Table 1 and Figs. 7C and D, the adsorption of both BSP-A and 4-CL on SF.CA followed the pseudo-first-order kinetic model more closely, with correlation coefficients exceeding 0.95. The strong agreement between calculated and experimental adsorption capacities (124.9 mg/g for BSP-A and 97.8 mg/g for 4-CL) reinforces the reliability of the pseudo-first-order model, suggesting that the process is primarily governed by physical adsorption mechanisms involving van der Waals and electrostatic forces^88,89.

Table 1.

The mathematical parameters of the kinetic models studied.

Model	Parameters	BSP-A	4-CL
Pseudo-First order	K1 (min-1(	0.0063	0.0052
	Qe (Cal) (mg/g)	124.9	97.81
	R2	0.98	0.96
	X2	0.88	1.34
Pseudo-Second order	k2 (g mg-1 min-1)	3.84 ×10-5	3.65× 10-5
	Qe (Cal) (mg/g)	153.9	124.7
	R2	0.95	0.94
	X2	2.001	2.29

Although the pseudo-first-order model offered the best overall fit, the pseudo-second-order model also showed reasonable consistency with the data. This partial conformity implies that, in addition to physisorption, weak chemical interactions—such as hydrogen bonding and surface complexation—may contribute to the overall adsorption process. Minor deviations from the first-order model are therefore attributed to the participation of surface oxygenated groups (e.g., –OH, –COOH, and –SO₃H) on SF.CA, which can form transient chemical bonds or donor–acceptor interactions with the aromatic rings of BSP-A and 4-CL^85,88. At higher pollutant concentrations, the adsorption process likely involves multi-layer accumulation. Initially, pollutant molecules form a monolayer through physical adsorption on the most reactive surface sites. As adsorption continues, additional molecules are retained via secondary interactions such as dipole–dipole forces or weak hydrogen bonding, leading to a marginal increase in adsorption capacity beyond the theoretical monolayer limit⁹⁰. The slight differences in adsorption performance between BSP-A and 4-CL are attributed to their molecular structures and the specific affinities of their functional groups toward the surface functionalities of SF.CA.

From a practical standpoint, the predominance of physisorption offers a key operational advantage. Because the process is largely reversible, the SF.CA adsorbent can be readily regenerated and reused with minimal structural degradation. The dominance of fast, surface-controlled van der Waals and electrostatic interactions ensures efficient and rapid uptake kinetics, confirming the suitability of SF.CA as a sustainable, high-capacity adsorbent for advanced water purification applications.

Starting concentration

The influence of the initial pollutant concentration on the adsorption efficiency of SF.CA toward bisphenol A (BSP-A) and 4-chlorophenol (4-CL) was systematically investigated over the concentration range of 50–400 mg/L. All experiments were performed under constant conditions, including a solution volume of 100 mL, a contact time of 24 h, an adsorbent dosage of 0.25 g/L, pH 8 for BSP-A and pH 7 for 4-CL, and temperatures of 303, 313, and 323 K. As illustrated in Figs. 8A–B, the adsorption capacity of SF.CA increased progressively with increasing pollutant concentration for both contaminants. At low concentrations, numerous vacant and energetically favorable sites on the SF.CA surface enabled rapid binding through electrostatic attraction and diffusion from the bulk solution. With increasing concentration, the stronger concentration gradient enhanced the driving force and diffusion rate of the pollutant molecules, promoting more frequent collisions with the adsorbent and consequently improving adsorption efficiency. Beyond a certain threshold, however, further increases in concentration produced only marginal gains in adsorption, indicating that the available active sites had become saturated and equilibrium had been reached⁹¹.

Fig. 7.

Contact-time effect and kinetic modeling of BSP-A and 4-CL adsorption by SF.CA: adsorption uptake profiles (A), intra-particle diffusion model (B), pseudo-first-order model (C), and pseudo-second-order model (D).

At equilibrium, the maximum adsorption capacities of SF.CA for BSP-A were 243.4 mg/g at 303 K, 231.5 mg/g at 313 K, and 223.8 mg/g at 323 K (Fig. 8A). For 4-CL, the corresponding capacities were 241.4 mg/g at 303 K, 223.6 mg/g at 313 K, and 190.3 mg/g at 323 K (Fig. 8B). The gradual decline in adsorption capacity with increasing temperature for both pollutants confirms that the adsorption process is exothermic in nature. At higher temperatures, the kinetic energy of solute molecules increases, weakening the physical interactions—such as van der Waals and electrostatic forces—responsible for adsorption and promoting partial desorption from the surface.

Although both pollutants displayed similar overall trends, their adsorption behaviors differed in magnitude and mechanism because of variations in molecular structure and interaction strength with the SF.CA surface. BSP-A consistently exhibited slightly higher adsorption capacities than 4-CL under identical conditions, a difference attributed to its two hydroxyl groups that enhance interactions with the aromatic and oxygenated functional groups of SF.CA. In contrast, 4-CL, which contains only one hydroxyl group and a chlorine substituent, is less polar and forms weaker hydrogen bonds. The electron-withdrawing –Cl group decreases the electron density of its aromatic ring, thereby weakening π–π stacking and electrostatic interactions with the carbonaceous surface^69,92,93.

Molecular size also plays a role in the observed behavior. The bulkier structure of BSP-A allows stronger π–π stacking with the graphitic domains of SF.CA but may hinder its diffusion into smaller pores, while the smaller 4-CL molecules diffuse more readily through the pore system, achieving near-comparable capacities despite their weaker interactions^94–96. Temperature-dependent changes further highlight mechanistic distinctions: both systems showed declining capacities with rising temperature, but the decrease was more pronounced for 4-CL, implying that its adsorption is governed mainly by weak physical forces that are more easily disrupted by thermal motion. BSP-A, by contrast, interacts through multiple non-covalent mechanisms that confer greater thermal stability to the adsorbed species.

Overall, these results demonstrate that BSP-A possesses a stronger adsorption affinity and higher thermal stability on SF.CA than 4-CL because of its multifunctional structure and capacity for multipoint interactions. Nevertheless, both pollutants exhibit high removal efficiencies under optimized conditions, confirming the versatility of SF.CA as an effective adsorbent for phenolic contaminants. The dependence of adsorption on the starting concentration follows a typical saturation pattern, with rapid uptake at low concentrations followed by a plateau as surface sites become occupied. The exothermic nature of the process further indicates that adsorption is thermodynamically favored at lower temperatures.

Classic isotherm models

Understanding adsorption equilibrium is fundamental for evaluating the efficiency, selectivity, and capacity of an adsorbent toward specific pollutants in aqueous systems. Equilibrium isotherm models describe the distribution of adsorbate molecules between the solid and liquid phases at constant temperature and provide important information on surface properties, adsorption mechanisms, and the degree of surface heterogeneity^88,97. To clarify the adsorption behavior of BSP-A and 4-CL on SF.CA, three classical isotherm models—Langmuir, Freundlich, and Dubinin–Radushkevich (D–R)—were applied to the equilibrium data (Fig. 8E–P). The Langmuir model assumes monolayer adsorption on energetically equivalent surface sites, where each site accommodates one adsorbate molecule^89,97. In contrast, the Freundlich model describes adsorption on heterogeneous surfaces containing active sites with different adsorption energies⁸⁵. Non-linear regression analysis was used to fit the experimental data, and the fitting reliability was evaluated using the correlation coefficient (R²) and chi-squared (χ²) values.

The equilibrium data showed good agreement with both the Langmuir and Freundlich models (Table 2, Fig. 8C–F), indicating that BSP-A and 4-CL adsorption onto SF.CA cannot be described by a single idealized surface behavior. Instead, the adsorption process involves both monolayer occupation of accessible active sites and heterogeneous interactions associated with the chemically diverse surface of sulfonated coal. The good agreement with the Langmuir model confirms that the adsorption process approaches surface saturation at equilibrium, meaning that uptake becomes limited once the available active sites are progressively occupied^89,97. Meanwhile, the Freundlich contribution reflects the non-uniform distribution of oxygenated and sulfonated groups, surface defects, and microstructural irregularities generated during coal modification.

Table 2.

the mathematical characteristics of the traditional isotherm models under consideration.

Models	Parameters	BSP-A			4-CL
		303 K	303 K	323 K	303 K	303 K	323 K
Langmuir	Qmax(mg/g)	246.24	230.98	224.83	243.01	224.04	189.87
	b(L/mg)	1.02×10-6	6.21×10-8	5.28×10-8	1.83×10-8	2.33×10-8	1.23×10-8
	R2	0.998	0.997	0.998	0.999	0.997	0.995
	X2	0.138	0.219	0.116	0.041	0.239	0.468
	RL	0.999	0.999	0.999	0.999	0.999	0.999
Freundlich	1/n	0.45	0.406	0.44	0.423	0.448	0.436
	kF (mg/g)	257.2	241.42	240.02	259.36	243.46	205.63
	R2	0.998	0.999	0.998	0.997	0.998	0.999
	X2	0.132	0.062	0.134	0.324	0.136	0.091
D-R	β (mol2/kJ2)	1.474	1.87	2.33	2.42	2.83	3.03
	Qm (mg/g)	256.71	250.001	242.93	266.12	244.63	209.04
	E (kJ/mol)	0.582	0.517	0.463	0.455	0.420	0.406
	R2	0.998	0.997	0.998	0.997	0.998	0.999
	X2	0.117	0.191	0.127	0.342	0.120	0.091
Monolayer model of one energy site	R2	0.999	0.999	0.999	0.999	0.997	0.995
	X2	0.03	0.04	0.08	0.041	0.239	0.468
	n	3.21	3.81	3.74	3.97	3.84	3.96
	Nm (mg/g)	76.75	60.59	60.09	61.07	58.28	47.90
	Qsat (mg/g)	245.78	230.84	224.73	242.44	223.79	188.72
	C1/2 (mg/L)	72.88	77.74	88.08	87.97	96.66	99.001
	ΔE (kJ/mol)	- 4.21	-4.38	-4.47	-14.42	-14.66	-15.06

The adsorption of 4-CL showed slightly stronger agreement with the Langmuir model, suggesting that its uptake occurs mainly through monolayer coverage on relatively uniform adsorption sites. This behavior may be related to the smaller molecular size of 4-CL, which facilitates access to active sites and promotes more regular surface coverage. In contrast, BSP-A showed closer agreement with the Freundlich model, indicating a stronger contribution from surface heterogeneity. This can be attributed to the larger molecular structure of BSP-A and the presence of two phenolic hydroxyl groups, which allow simultaneous interaction with different surface sites.

The Freundlich heterogeneity constants (1/n) for BSP-A were in the range of 0.40–0.45, confirming favorable adsorption on a moderately heterogeneous surface. Since 1/n values between 0 and 1 indicate favorable adsorption, these results demonstrate that SF.CA possesses active sites with different affinities toward phenolic molecules⁹⁷. The relatively high Freundlich capacity constants also support the strong uptake ability of SF.CA. However, the high R² values obtained from the Langmuir model indicate that monolayer adsorption remains a major contribution, while multilayer accumulation becomes less dominant at higher concentrations⁹⁷.

The dimensionless Langmuir separation factor, , was calculated using , where b is the Langmuir affinity constant and is the initial pollutant concentration. The obtained values were below unity for both pollutants, confirming favorable adsorption onto SF.CA across the studied concentration range⁹². Although the fitted Langmuir b values were small, the resulting values close to unity are mathematically consistent with the Langmuir separation-factor equation because the product remains much lower than 1 under the investigated conditions. Therefore, should be interpreted as favorable adsorption occurring close to the linear Langmuir limit. In this context, b reflects the apparent Langmuir affinity constant, whereas reflects the maximum adsorption capacity associated with the number of accessible adsorption sites. Thus, low b values can coexist with high values when the adsorbent provides abundant adsorption sites, but the affinity term remains low within the studied concentration range.

Comparable Langmuir behavior has been reported in several phenolic-pollutant adsorption systems, supporting the interpretation of high adsorption capacity with favorable values. Tazik et al.⁷⁹ reported 4-chlorophenol adsorption onto amine-functionalized activated carbon with a high maximum adsorption capacity of 316.1 mg/g and values within the favorable range of 0–1. Fuzil et al.⁹⁵ reported BSP-A adsorption onto alginate and graphene oxide–alginate beads, where the Langmuir values were also between 0 and 1, while the maximum adsorption capacities reached 250.00 and 384.62 mg/g, respectively. Similarly, Siva Kumar et al⁹⁸. reported Langmuir-fitted adsorption of 2,4,6-trichlorophenol onto pinecone biomass with mg/g, b=0.009 L/mg, and , while confirming that indicates favorable adsorption and represents linear adsorption. Zbair et al⁹⁹. also reported high-capacity BSP-A adsorption onto activated carbon prepared from argan nutshell, with Langmuir reaching 1250 mg/g. These studies support that high Langmuir capacities and favorable values are commonly observed for phenolic-pollutant adsorption systems. However, the absolute magnitude of b is dependent on the adsorbate–adsorbent system, concentration range, fitting form, and unit convention; therefore, the present low b values are interpreted together with the calculated , , and experimental uptake rather than by direct comparison of b alone.

The Langmuir maximum adsorption capacities were consistent with the experimental uptake values. For BSP-A, the values were 246.2, 230.9, and 224.8 mg/g at 303, 313, and 323 K, respectively. For 4-CL, the corresponding values were 243.0, 224.0, and 189.8 mg/g. These high capacities confirm the strong adsorption performance of SF.CA toward both phenolic pollutants. The gradual decrease in with increasing temperature indicates that adsorption becomes less favorable at elevated temperature, supporting an exothermic adsorption mechanism.

Further insight into the energetic nature of the adsorption process was obtained using the Dubinin–Radushkevich (D–R) model, which evaluates the mean free adsorption energy (E) and helps distinguish between physical and chemical adsorption mechanisms (Fig. 8G and H)^6,28. The E values calculated from were 0.582, 0.517, and 0.463 kJ mol⁻¹ for BSP-A at 303, 313, and 323 K, respectively, and 0.455, 0.420, and 0.406 kJ mol⁻¹ for 4-CL at the same temperatures (Table 2). These values are lower than 8 kJ mol⁻¹, indicating that adsorption proceeds mainly through a physisorption-dominated mechanism rather than ion exchange or strong chemisorption^6,28. According to the D–R classification, E values below 8 kJ mol⁻¹ are typically associated with physical adsorption controlled by weak non-covalent interactions, while values in the range of 8–16 kJ mol⁻¹ indicate ion-exchange or mixed physicochemical adsorption^6,28. The low E values suggest that BSP-A and 4-CL adsorption onto SF.CA is mainly governed by physical mechanisms such as van der Waals forces and electrostatic attractions.

Although the D–R energy values confirm the predominance of physisorption, the presence of –SO₃H, –OH, –COOH, and C=O groups on the modified coal surface can strengthen pollutant retention through multiple non-covalent interactions. This interaction network explains the high adsorption capacities while maintaining the reversible character required for adsorbent regeneration and reuse. The consistency between the experimental data and the applied isotherm models confirms that SF.CA behaves as an efficient and high-capacity adsorbent for removing phenolic compounds such as BSP-A and 4-CL from aqueous media.

Advanced isotherm models

To gain a deeper understanding of the equilibrium behavior and molecular mechanisms governing adsorption, a statistical physics isotherm model was employed to complement the classical models. This advanced approach provides a more comprehensive interpretation of adsorption phenomena by accounting for both steric (structural) and energetic (thermodynamic) factors. Unlike empirical models that merely describe adsorption capacity, the statistical physics model elucidates the molecular interactions between the adsorbate species (BSP-A and 4-CL) and the reactive surface functional groups of SF.CA that act as adsorption receptors. Through mathematical simulation, the model quantifies several steric parameters, including Nₘ, representing the total number of active sites occupied during adsorption; n, the number of pollutant molecules anchored per receptor site; and Q_sat, the theoretical saturation capacity corresponding to complete surface coverage of SF.CA by the adsorbed species. These steric descriptors are particularly significant because they reveal the spatial arrangement and accessibility of adsorption sites, as well as molecular orientation effects that influence adsorption efficiency. In addition to the structural parameters, the model incorporates energetic and thermodynamic quantities such as internal energy (E_int), entropy (Sa), adsorption energy (ΔE), and enthalpy (G). Together, these parameters provide valuable insights into the nature of the interactions, their spontaneity, and the driving forces behind the adsorption process. Negative ΔE and G values typically indicate exothermic and spontaneous adsorption, while positive or less negative entropy changes (Sa) reflect the degree of structural reorganization and disorder occurring at the solid–liquid interface during pollutant attachment.

To ensure the accuracy and reliability of the statistical physics analysis, the equilibrium data were fitted using non-linear regression based on the Levenberg–Marquardt algorithm with multivariable optimization (Table S1). Several theoretical adsorption configurations were tested, and the model with the best statistical agreement was selected. The single active-site monolayer model provided the most suitable fit for both BSP-A and 4-CL adsorption onto SF.CA (Fig. 8I and J; Table 2). The significance of this result is that the equilibrium uptake is mainly controlled by a dominant population of accessible receptor sites with comparable apparent adsorption energies.

Mechanistically, the fitted model suggests that BSP-A and 4-CL adsorption proceeds mainly through monolayer-like saturation of accessible sulfonated and oxygenated sites, with limited lateral interaction under the model assumptions. Structurally, this indicates that although sulfonation generates a heterogeneous surface containing defects, roughness, and different functional groups, the adsorption process is governed primarily by the most active sites, including –SO₃H, –OH, –COOH, and C=O groups. Therefore, the statistical physics model complements the classical Langmuir–Freundlich interpretation: Langmuir-type behavior reflects apparent saturation of dominant active sites, while Freundlich behavior reflects the heterogeneous nature of the SF.CA surface.

Steric properties

Number of adsorbed molecules per site (n)

The statistical physics steric parameter n, defined as the average number of adsorbate molecules associated with one active receptor site, provides a theoretical basis for interpreting the possible spatial arrangement of bisphenol A (BSP-A) and 4-chlorophenol (4-CL) on the SF.CA surface. According to the assumptions commonly used in statistical-physics adsorption models, n<1 is generally associated with a horizontal or parallel adsorption configuration, where one adsorbate molecule may interact with more than one receptor site. In contrast, n>1 suggests that more than one adsorbate molecule can be associated with a single active site, which is commonly interpreted as possible non-parallel, tilted, or vertical multi-molecular adsorption through adsorbate–adsorbate association near the same receptor site^86,100.

For SF.CA, the n values derived from the advanced statistical physics model ranged from 3.21 to 3.81 for BSP-A and from 3.84 to 3.97 for 4-CL across the studied temperature range of 303–323 K (Figs. 9A and B; Table 2). Based on the established interpretation of this parameter, these values indicate that several pollutant molecules may be associated with one active adsorption site. This behavior is consistent with a possible multi-molecular adsorption configuration promoted by the chemically active surface of SF.CA. However, these values do not independently prove molecular stacking; rather, they suggest that adsorbate–adsorbate association and cooperative binding may contribute to the overall adsorption process. The slightly higher n values obtained for 4-CL may be related to its smaller molecular size, which could facilitate closer packing and more efficient association near sulfonated and oxygenated active sites. In comparison, the slightly lower n values of BSP-A may reflect steric limitations caused by its larger molecular structure. The presence of –SO₃H, –COOH, and –OH groups on SF.CA can provide receptor sites for interactions, thereby supporting the model-derived possibility of multi-molecular association at the solid–liquid interface^43,101,102.

Fig. 8.

Equilibrium adsorption and isotherm modeling of BSP-A and 4-CL adsorption by SF.CA: initial-concentration effect (A, BSP-A; B, 4-CL), Langmuir model (C, BSP-A; D, 4-CL), Freundlich model (E, BSP-A; F, 4-CL), Dubinin–Radushkevich model (G, BSP-A; H, 4-CL), and single-site monolayer statistical-physics model (I, BSP-A; J, 4-CL).

Temperature variations also affected the calculated n parameter. Between 303 and 313 K, n increased slightly for both pollutants, suggesting, improved molecular mobility and more effective occupation or association around active sites^100,103. This may be related to enhanced diffusion and greater accessibility of pollutant molecules to the functionalized surface. However, increasing the temperature further to 323 K caused a slight decrease in n, indicating that excessive thermal motion may reduce the stability of adsorbate association near the receptor sites and weaken the weak interactions responsible for retaining the adsorbed molecules. This trend is consistent with the exothermic nature of adsorption and with the temperature sensitivity of weak physical interactions^104–107.

Density of the active sites (Nm)

Quantitative assessment of the active adsorption sites (Nₘ) provides a crucial measure of the total number of functional centers effectively participating in pollutant uptake on the SF.CA surface (Fig. 9G–H). For bisphenol A (BSP-A), the calculated Nₘ values were 76.75 mg/g at 303 K, 60.59 mg/g at 313 K, and 60.09 mg/g at 323 K, whereas for 4-chlorophenol (4-CL), the corresponding values were 61.07 mg/g, 58.28 mg/g, and 47.90 mg/g, respectively (Fig. 9C and D; Table. 2). The comparatively higher Nₘ values observed for BSP-A confirm that a greater number of reactive sites were effectively occupied, reflecting its stronger affinity and superior uptake performance relative to 4-CL. This difference can be attributed to the dual hydroxyl functionalities in BSP-A, which favor multiple hydrogen-bonding interactions with the surface –SO₃H, –COOH, and –OH groups of SF.CA, while 4-CL, bearing only one hydroxyl group and a chlorine substituent, interacts less extensively through weaker π–π and dipolar forces^69,92,93.

A clear temperature-dependent trend was also evident for both pollutants, with Nₘ decreasing as the temperature increased from 303 K to 323 K. This decline indicates that elevated thermal conditions adversely influence the stability and availability of reactive adsorption sites, possibly leading to partial deactivation or weakening of the adsorbate–adsorbent interactions. At higher temperatures, enhanced molecular mobility and decreased solution viscosity accelerate the diffusion of BSP-A and 4-CL molecules, shortening their residence time near the adsorption interface and thus limiting effective anchoring at surface sites. Additionally, the increased kinetic energy of the adsorbed molecules promotes desorption and reduces the surface saturation threshold, consistent with the exothermic nature of the adsorption process^108,109.

Overall, the Nₘ values demonstrate that adsorption on SF.CA proceeds via a temperature-sensitive, site-specific mechanism, where the number of active functional centers and their effective participation are governed by both surface chemistry and thermal dynamics. The higher density of active sites engaged in BSP-A removal confirms that sulfonation successfully enhances surface reactivity and active-site accessibility, thereby facilitating efficient sequestration of multifunctional phenolic pollutants even under moderately elevated temperatures.

Adsorption capacity at the saturation state of (Q_sat)

The fully saturated adsorption capacity (Q_sat) represents the ultimate measure of an adsorbent’s efficiency and structural stability under complete surface coverage conditions. In the case of SF.CA, Q_sat values provide direct evidence of its high affinity and robust capability for sequestering phenolic pollutants such as bisphenol A (BSP-A) and 4-chlorophenol (4-CL) (Fig. 9E and F; Table. 2). The two major factors determining Q_sat are the number of active adsorption sites (Nₘ) and the number of pollutant molecules attached per site (n). These parameters jointly define the maximum adsorption density achievable on the SF.CA surface. For BSP-A, Q_sat values were recorded as 245.78 mg/g at 303 K, 230.80 mg/g at 313 K, and 224.70 mg/g at 323 K, while for 4-CL, the corresponding capacities were 242.40, 223.80, and 188.70 mg/g, respectively. The relatively higher Q_sat of BSP-A across all temperatures indicates a stronger overall binding affinity and more effective utilization of the available reactive sites. This can be attributed to its molecular structure, which contains two hydroxyl groups capable of forming multiple hydrogen bonds with the –SO₃H, –COOH, and –OH functionalities on the SF.CA surface, enhancing both molecular anchoring and surface coverage. In contrast, 4-CL, which possesses only one hydroxyl group and a chlorine substituent, forms fewer hydrogen-bonding interactions and relies mainly on π–π stacking and dipole–dipole forces, leading to comparatively lower saturation capacities.

The inverse correlation between Q_sat and temperature (Fig. 9E and F) clearly confirms that adsorption is an exothermic process. As temperature increases, molecular agitation disrupts weak van der Waals and hydrogen-bonding interactions, thereby promoting desorption and reducing equilibrium coverage. This thermal dependence mirrors the observed trend for Nₘ, implying that the total number of active sites exerts a greater influence on adsorption performance than the degree of multi-molecular occupancy per site. Such temperature sensitivity also highlights the predominance of physical adsorption mechanisms, in which thermal fluctuations directly impact molecular stability at the solid–liquid interface. Collectively, these results substantiate that SF.CA maintains a high adsorption capacity and structural integrity even under varying thermal conditions. The high Q_sat values, combined with consistent site accessibility and moderate energy requirements, confirm that the sulfonation process effectively optimizes both surface functionality and textural properties. Consequently, SF.CA demonstrates strong potential as a cost-effective and thermally stable adsorbent for large-scale removal of phenolic contaminants from aqueous environments.

Energetic properties

Adsorption of energy and mechanism

Evaluating the adsorption energy (ΔE) provides an essential diagnostic for understanding the fundamental nature of adsorbate–adsorbent interactions and differentiating between physical and chemical adsorption mechanisms. In general, ΔE values lower than 40 kJ/mol correspond to physisorption processes dominated by non-covalent forces, while values exceeding 80 kJ/mol indicate chemisorption, where covalent or coordinative bonds are formed. Within the low-energy regime, several weak interactions collectively sustain adsorption: hydrogen bonding (< 30 kJ/mol), dipole–dipole coupling (2–29 kJ/mol), coordination exchange (~ 40 kJ/mol), van der Waals dispersion forces (4–10 kJ/mol), electrostatic attraction (2–50 kJ/mol), and hydrophobic association (~ 5 kJ/mol). These energy thresholds are widely validated in surface-science literature and serve as reliable benchmarks for interpreting pollutant–surface interactions in aqueous systems¹¹⁰. The adsorption energies for bisphenol A (BSP-A) and 4-chlorophenol (4-CL) were calculated using Equation (5), which incorporates the solubility (S) of each compound, the gas constant (R = 0.008314 kJ mol⁻¹ K⁻¹), the system temperature (T), and the pollutant concentration at half-saturation conditions (C₁/₂):

5

The estimated ΔE values were below 5 kJ/mol for BSP-A and below 16 kJ/mol for 4-CL (Figs. 9K–L; Table 2), confirming that adsorption on SF.CA proceeds primarily through physisorption. These magnitudes are well below the typical threshold for chemical bonding, indicating that no covalent or coordination bonds are formed between pollutant molecules and the sulfonated coal framework. The negative sign of ΔE further verifies the exothermic nature of adsorption, signifying that the process is thermodynamically spontaneous and energetically favorable^107,110. The low energy of interaction is consistent with the presence of abundant polar functional groups (–SO₃H, –COOH, and –OH) on SF.CA, which promote strong hydrogen bonding and electrostatic attractions with the polar hydroxyl groups of BSP-A and 4-CL. Hydrogen bonds form between the oxygen atoms of SF.CA’s oxygenated sites and the hydrogen atoms of the adsorbate hydroxyl groups, creating a network of directional interactions that stabilize the adsorption complex. Such bonds typically contribute 10–30 kJ/mol, matching the energetic range observed here. The negatively charged surface of SF.CA—originating from sulfonic and carboxylate moieties—also enhances electrostatic attraction with the partially positive hydrogen atoms of the pollutants, providing additional stabilization and explaining the spontaneous, exothermic nature of the process^{103,109,111–114}

The substantial dipole moments of both BSP-A and 4-CL facilitate alignment with the polar domains of SF.CA, allowing dipole–dipole interactions to complement hydrogen bonding and van der Waals forces. Notably, the relatively higher ΔE for 4-CL (≈ 15 kJ/mol) compared with BSP-A (≈ 5 kJ/mol) implies that dipole–dipole and electrostatic effects play a more significant role in 4-CL adsorption, whereas BSP-A—owing to its bulkier structure and two hydroxyl groups—interacts through weaker, more delocalized van der Waals and hydrogen-bond networks distributed over multiple sites. Consequently, BSP-A tends to form multilayer physical assemblies with relatively lower individual bond energies but higher overall adsorption^113,114.

These energetic distinctions reflect intrinsic differences in molecular geometry, polarity, and electron distribution. BSP-A, possessing two phenolic rings and multiple polar sites, exhibits strong π–π stacking and hydrogen-bonding synergy with the aromatic and oxygenated surface domains of SF.CA. This leads to enhanced surface coverage and easier desorption upon thermal activation—an advantageous feature for adsorbent regeneration. Conversely, the smaller 4-CL molecule, containing only one hydroxyl and one chlorine substituent, forms fewer hydrogen bonds but engages in stronger local dipole–dipole coupling due to its asymmetric charge distribution. The presence of the electron-withdrawing –Cl group increases molecular polarity and enhances electrostatic attraction toward the negatively charged surface, explaining its slightly higher ΔE yet lower total uptake.

Overall, the ΔE values confirm that BSP-A and 4-CL adsorption onto SF.CA is a predominantly physical, exothermic, and spontaneous process governed by a synergistic interplay of hydrogen bonding, electrostatic forces, dipole–dipole interactions and van der Waals attractions. The low energy barrier facilitates rapid adsorption–desorption reversibility, ensuring efficient pollutant removal while maintaining structural integrity of the adsorbent. These findings highlight the high thermodynamic stability, reusability, and environmental suitability of SF.CA as a sustainable adsorbent for phenolic contaminants under ambient conditions.

Thermodynamic functions

Entropy

The entropy (Sₐ) characteristics associated with the adsorption of bisphenol A (BSP-A) and 4-chlorophenol (4-CL) on the sulfonated coal (SF.CA) surface offer essential insights into the degree of molecular order and configurational rearrangement occurring at the solid–liquid interface under different concentrations and temperatures. The quantitative evaluation of Sₐ was conducted using Equation (6), which incorporates the previously determined steric parameters ( and n) and the equilibrium pollutant concentration corresponding to half-saturation (C₁/₂):

6

The entropy profiles reveal a marked decline in upon adsorption of both BSP-A and 4-CL molecules, particularly at elevated solute concentrations (Figs. 10A–D). This consistent reduction reflects a significant decrease in the configurational randomness and freedom of movement of the adsorbed molecules and the surface functional groups after adsorption. As pollutant molecules occupy the available active sites, the system transitions from a relatively disordered aqueous phase to a more ordered solid–liquid interfacial state. The pronounced entropy drop therefore indicates restricted molecular mobility, enhanced interfacial organization, and formation of stable adsorbate–adsorbent complexes, confirming that adsorption proceeds spontaneously but with an associated loss of randomness in the system^107,109.

Fig. 9.

Statistical-physics-derived steric and thermodynamic parameters for BSP-A and 4-CL adsorption by SF.CA: number of adsorbed molecules per site (A, BSP-A; B, 4-CL), active-site density (C, BSP-A; D, 4-CL), saturation adsorption capacity (E, BSP-A; F, 4-CL), entropy (G, BSP-A; H, 4-CL), free enthalpy (I, BSP-A; J, 4-CL), and internal energy (K, BSP-A; L, 4-CL).

The maximum entropy values for BSP-A adsorption were recorded at equilibrium concentrations of 69.9 mg/L (303 K), 72.8 mg/L (313 K), and 77.3 mg/L (323 K), whereas for 4-CL, the corresponding values were 76.9, 80.1, and 82.3 mg/L, respectively (Figs. 9G and H). These equilibrium concentrations are consistent with those observed at the half-saturation states, confirming that adsorption is thermodynamically controlled by the occupation of energetically favorable sites. Beyond this point, the decline in entropy reflects the progressive exhaustion of high-energy binding centers and the transition toward an ordered multilayer or surface-saturated state. The observed temperature-dependent behavior suggests that entropy values slightly increase with rising temperature due to the enhanced kinetic energy and diffusion of adsorbate molecules, which facilitate access to deeper or less accessible pores before eventual stabilization. However, the overall negative change in Sₐ across the concentration range indicates that adsorption of both BSP-A and 4-CL on SF.CA is an entropically unfavorable but energetically favorable process. This aligns with a typical physisorption mechanism, where spontaneous adsorption is driven primarily by exothermic enthalpic contributions rather than by entropy gain^110,115.

In summary, the entropy analysis demonstrates that the adsorption of phenolic pollutants onto SF.CA induces a transformation toward a more ordered surface configuration. The substantial reduction in system disorder reflects the formation of structured molecular assemblies, restricted rotational and translational mobility, and high affinity of adsorbates toward functionalized adsorption sites, thereby confirming the stability and efficiency of SF.CA as a thermodynamically robust adsorbent for BSP-A and 4-CL removal.

Internal energy and free enthalpy

The thermodynamic investigation of the internal energy (E₍ᵢₙₜ₎) and free enthalpy (G) associated with the adsorption of bisphenol A (BSP-A) and 4-chlorophenol (4-CL) onto sulfonated coal (SF.CA) provides valuable insights into the spontaneity, energetic feasibility, and temperature dependence of the removal process. These parameters were computed as functions of pollutant concentration and system temperature using Equations (7) and (8), which incorporate the steric parameters ( and n), the half-saturation concentration (C₁/₂), and the translational partition function ()¹⁰²:

7

8

The E_int values for BSP-A and 4-CL adsorption were consistently negative across all studied conditions (Figs. 9I and J), confirming that the adsorption process is exothermic and thermodynamically favorable. A gradual decline in E_int magnitude was observed with increasing temperature (from 303 K to 323 K), suggesting that elevated thermal energy weakens the physical interactions between adsorbate molecules and active sites on SF.CA. This decrease in internal energy reflects the temperature-induced reduction in adsorption affinity, which is typical of physisorption processes driven by weak intermolecular forces. The consistent negativity of E_int indicates that adsorption proceeds spontaneously with energy release, promoting the stabilization of the pollutant–adsorbent system. The reduction in E_int at higher temperatures may also suggest a transition toward a more energetically stable configuration, where adsorbed molecules rearranged to minimize system energy. This behavior aligns with the physical adsorption mechanism, in which exothermic energy release accompanies surface–molecule interactions without the formation of strong covalent bonds.

The enthalpy (G) trends mirrored those of internal energy (Figs. 9K and L), further reinforcing the exothermic and spontaneous nature of the adsorption process. Negative G values across all conditions confirm that adsorption occurs naturally without external energy input and that the system attains lower energy states upon pollutant binding^110,115. The observed reduction in G magnitude with increasing temperature highlights the decreasing thermodynamic feasibility of adsorption at elevated temperatures, consistent with the partial desorption or diminished surface interaction strength observed in exothermic physical adsorption systems. Overall, the negative internal energy and enthalpy values validate that BSP-A and 4-CL adsorption onto SF.CA is an exothermic, spontaneous, and enthalpy-driven process^110,116. The temperature-dependent decline in both parameters indicates that adsorption occurs primarily via weak physical forces, confirming the physisorption mechanism inferred from the energetic and steric analyses. These results demonstrate the high thermodynamic stability, efficiency, and reusability of SF.CA as a robust adsorbent for the sustainable removal of phenolic pollutants from aqueous environments.

Interaction mechanism and adsorption pathways of sulfonated coal

The sulfonation of coal induces profound structural and chemical modifications that markedly enhance its adsorption efficiency toward organic pollutants. The modification primarily proceeds through an electrophilic aromatic substitution mechanism, in which electrophilic SO₃H⁺ species generated from concentrated p-xylene-2-sulfonic acid attack the aromatic domains of the coal matrix. This process leads to the covalent incorporation of aryl–sulfonic groups (–SO₃H), introducing strong Brønsted acid sites capable of proton donation and ion-exchange interactions (Fig. 10)^24,117. In parallel, the highly acidic environment promotes oxidative activation at defect sites and aromatic edges, producing additional oxygen-bearing functional groups such as carboxyl (–COOH), hydroxyl (–OH), and carbonyl (C=O). These oxygenated groups synergistically increase the surface polarity, reactivity, and density of active adsorption centers^38,44,118.

Fig. 10.

Proposed schematic representation of raw coal functionalization with PXSA and the adsorption mechanisms of BSP-A and 4-CL on SF.CA.

The combined incorporation of –SO₃H and oxygenated functionalities produces pronounced physicochemical transformations in the coal surface (Fig. 10). The total acidity and surface-active site density increase substantially, while the point of zero charge (pHₚ_zc) shifts to lower values due to the strong acidity of –SO₃H and –COOH moieties. This shift imparts a more negative surface charge under neutral pH conditions, thereby enhancing electrostatic attraction between the adsorbent and positively charged or protonated pollutant species. Moreover, the increased polarity and hydrophilicity improve surface wettability and particle dispersion in aqueous environments, facilitating better contact between pollutants and the adsorbent. This improved interface accessibility effectively minimizes external mass-transfer resistance, thereby accelerating adsorption kinetics and increasing overall uptake efficiency^{24,38,119,120}.

These chemical and structural transformations translate directly into improved adsorption performance toward phenolic pollutants, particularly bisphenol A (BSP-A) and 4-chlorophenol (4-CL) (Fig. 10). The strong Brønsted acidity of –SO₃H groups promotes acid–base interactions, while the –OH and –COOH groups participate as hydrogen bond donors and acceptors with the hydroxyl moieties of phenolic molecules. Furthermore, the aromatic backbone of coal enables π–π stacking and electron donor–acceptor (EDA) interactions with the aromatic rings of BSP-A and 4-CL, enhancing molecular-level affinity. The negatively charged surface also favors electrostatic interactions with partially protonated phenolic species under near-neutral conditions, while van der Waals dispersion forces Fig. 10 between the aromatic domains of coal and the nonpolar regions of the adsorbates provide additional stabilization^{106,121–123}.

The synergistic interplay of these mechanisms—acid–base binding, hydrogen bonding, π–π stacking, electrostatic attraction, and dispersion forces—constitutes the basis for the superior adsorption efficiency of sulfonated coal (Fig. 10)^105,120,121. Rather than being dominated by a single mechanism, adsorption arises from the collective and complementary contributions of multiple weak interactions, which together yield high pollutant uptake capacities and rapid adsorption kinetics. Enhanced hydrophilicity further reduces interfacial resistance, while the high density of active functional sites ensures effective pollutant binding per unit mass of adsorbent^{31,47,122,123}. In summary, sulfonated coal functions as a multifunctional, sustainable, and cost-effective adsorbent, exhibiting strong affinity for phenolic and other aromatic contaminants in aqueous systems. Its tunable surface chemistry—controlled by the balance of acidic and oxygenated functionalities—provides a rational pathway for designing next-generation carbonaceous materials for environmental remediation, advanced wastewater treatment, and catalytic applications.

Recyclability

The ability of an adsorbent to be regenerated and reused without significant loss of efficiency is a key determinant of its practical applicability, cost-effectiveness, and environmental sustainability in wastewater treatment operations. The sulfonated coal (SF.CA) exhibited excellent recyclability and structural stability, maintaining high adsorption capacities for bisphenol A (BSP-A) and 4-chlorophenol (4-CL) over five consecutive adsorption–desorption cycles. Regeneration was accomplished by washing the spent SF.CA particles with distilled water, followed by drying at 50 °C for 8 h to eliminate residual moisture and restore active surface sites. Experimental results demonstrated that SF.CA retained over 80% of its initial adsorption capacity after five regeneration cycles, underscoring its robust performance and physicochemical stability. For BSP-A, the adsorption capacity remained above 240 mg/g during the first two cycles, gradually decreased to 232 mg/g after four cycles, and stabilized at 218 mg/g after five cycles. Similarly, for 4-CL, capacities exceeded 236 mg/g during the first two cycles, decreased to 223 mg/g after the fourth cycle, and reached 209 mg/g after the fifth. These results confirm that SF.CA maintains high pollutant removal efficiency even after repeated regeneration, reflecting the reversible nature of adsorption and the durability of surface functionalities. Future tests can investigate desorption efficiency, mass balance, and post-cycling FTIR/BET stability to optimize SF.CA regeneration.

The gradual decline in performance is attributed to partial saturation or blockage of adsorption sites, minor textural rearrangements, and potential surface oxidation or weakening of secondary interactions during repetitive drying and rehydration. Such changes can slightly reduce the accessibility of active sites or alter their polarity. Nevertheless, the preservation of key functional groups (–SO₃H, –COOH, and –OH) ensures that the dominant adsorption mechanisms—remain largely unaffected throughout multiple cycles. The strong recyclability of SF.CA highlights its mechanical and chemical resilience, as well as the dominance of physisorption mechanisms, which allow facile desorption of bound molecules without structural degradation. The water-based regeneration approach employed in this study further enhances the sustainability of the process by avoiding chemical solvents, minimizing secondary pollution, and reducing regeneration costs. From a practical standpoint, the high reusability and low-cost regeneration protocol position SF.CA as a viable, eco-efficient alternative to conventional single-use adsorbents for industrial wastewater treatment and large-scale environmental remediation. Its combination of high adsorption capacity, structural stability, and excellent regeneration potential establishes SF.CA as a promising sustainable material for long-term pollutant removal applications.

Comparative studies

A comparative evaluation was conducted to benchmark the adsorption performance of sulfonated coal (SF.CA) against a range of previously reported adsorbents for the removal of bisphenol A (BSP-A) and 4-chlorophenol (4-CL) (Table S2). The analysis demonstrates that SF.CA exhibits superior adsorption capacities, outperforming several commercial, nanostructured, and bio-based materials, thereby confirming its potential for large-scale environmental remediation applications. The maximum adsorption capacities (Q_max) of SF.CA were determined as 246.2 mg/g for BSP-A and 243 mg/g for 4-CL, values that are substantially higher than those reported for most conventional adsorbents. For comparison, activated carbon, one of the most widely used commercial adsorbents, displayed a significantly lower capacity for BSP-A (158.7 mg/g). Similarly, SF.CA outperformed nanoporous graphene, nanosized activated carbons, and amine-functionalized activated carbons (typically ~316.1 mg/g for 4-CL), confirming its strong affinity and high surface utilization efficiency toward phenolic compounds.

Furthermore, SF.CA showed marked superiority over advanced carbon-based nanomaterials, including carbon nanofibers (63.3 mg/g for BSP-A) and graphene oxide (137 mg/g for BSP-A), indicating that its optimized surface chemistry and hierarchical pore structure enhance molecular adsorption beyond that achieved by even high-cost engineered nanomaterials. In comparison with bio-derived adsorbents, such as pinecone-activated ZnCl₂ (73.5 mg/g) and alkali-modified biochar (71.4 mg/g), SF.CA exhibited adsorption capacities approximately three to four times higher. These differences can be attributed to the combined influence of sulfonation-induced acidity, oxygenated surface functionalities, and improved wettability, which collectively facilitate stronger adsorption for the phenolic pollutants. In addition, SF.CA significantly outperformed functionalized mineral-based adsorbents such as organoclays (94.2 mg/g) and bentonite–chitosan composites (27.4 mg/g). These findings confirm that the hybrid carbonaceous–sulfonated structure of SF.CA provides a more versatile and chemically active adsorption interface, capable of capturing phenolic contaminants with both high capacity and stability.

The comparative assessment clearly establishes SF.CA as a high-performance and sustainable adsorbent for the remediation of phenolic wastewater. Its combination of low-cost synthesis, chemical tunability, and reusability make it a promising alternative to traditional activated carbon and advanced nanomaterials. From an engineering perspective, these properties underscore the potential of SF.CA for scalable implementation in continuous-flow adsorption systems, particularly in industrial wastewater treatment, where efficiency, regenerability, and cost-effectiveness are paramount.

Conclusions

Direct sulfonation of the native coal using p-xylene-2-sulfonic acid hydrate (2,5-dimethylbenzenesulfonic acid; C₈H₁₀O₃S·xH₂O; CAS No. 609-54-1) proved to be a decisive modification route that changed not only the chemistry, but also the structural identity of the surface. The suppression of the (10) turbostratic band confirmed that the semi-crystalline graphitic ordering was broken and replaced by an amorphous, defect-rich architecture — this is a structural transformation that does not occur in conventional “carbonize → then sulfonate” sequences. This shifted surface state was reflected directly in the adsorption performance. The SF.CA material reached 245.78 mg/g for Bisphenol A and 242.44 mg/g for 4-chlorophenol, positioning the material among the most effective adsorbents for phenolic compounds.

Importantly, the adsorption behaviour was not only high — it was mechanistically defined. The single-site monolayer model provided the best statistical match (R² = 0.999; χ² down to 0.03), and the statistical-physics parameters (n = 3.21–3.97 and Nₘ = 76.75 to 47.90 mg/g) showed that adsorption took place through multi-molecular packing at each active site. The negative ΔE values (as low as −15.06 kJ/mol) confirmed that this packing was spontaneous and occurred through low-energy, exothermic interactions — consistent with the cooperative π–π stacking and hydrogen-bonding behaviour inferred from the steric and energetic interpretation.

These results verify that direct sulfonation is not simply a shorter preparation step, but an effective route that generates a surface capable of highly coordinated molecular assembly with quantifiable energetic and steric signatures. This elevates raw coal from a low-value precursor to a mechanistically predictable high-performance adsorbent. The present study is limited to batch conditions, two phenolic species, and the temperatures examined; therefore, regeneration, column performance, and real wastewater testing will be important future steps.

Author contributions

1.Gehad Hussein: Conceptualization, Methodology, Resources, Visualization, Data curation, Formal analysis, Writing -original draft, Writing – review & editing, 2. Eman S. H. Khaled: Supervision, Resources, Writing - original draft, 3.Ahmed A. Allam: validation, Writing - original draft, Writing – review & editing 4.Reham A. Mohamed: Supervision, Data curation, validation, Writing -original draft 5.Mabrouk Sami: Resources, validation, Writing -original draft, Writing – review & editing 6.May Bin-Jumah: Data curation, Funding, validation, Writing -original draft 7.Mostafa R. Abukhadra: Conceptualization, Methodology, Resources, Visualization, investigation, validation, Data curation, Formal analysis, Writing -original draft, Writing – review & editing, 8.Stefano Bellucci: Analysis, funding, validation, Writing – review & editing 9. Douaa Fathy: Resources, validation, Writing -original draft, Writing – review & editing.

Funding

The authors acknowledge Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2026R737), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia. This work is supported in part by the Italian Ministry for University and Research (MUR) with the resources of the FISA - Fondo Italiano per le Scienze Applicate, for the year 2023, project FISA-2023-00229 “Advanced Composite nanomaterials for WAter Management”.

Data availability

The data will be available based on reasonable request.

Declarations

Competing interest

The authors declare that they have no known competing or conflicting interests.

AI tools

AI tools were used during manuscript preparation for language editing and refinement of schematic figures under the authors’ supervision, who reviewed all outputs and take full responsibility for the final content.