Remediation of Caffeine from Aqueous Solutions Using Waste-Derived Adsorbents: A Polyaniline/Cuttlefish Bone Nanocomposite for Pollutant Removal

Abstract

Caffeine is commonly used in pharmaceutical and personal care products, where it serves both therapeutic and cosmetic purposes. However, its widespread presence in wastewater from the pharmaceutical and cosmetic industries has raised concerns about environmental contamination. This study explores the use of a polyaniline (PANI)/cuttlefish bone (CB) nanocomposite as an effective adsorbent for the removal of caffeine from aqueous solutions. The nanocomposite was synthesized by incorporating polyaniline (PANI) onto cuttlefish bone (CB) flakes, resulting in a material with a hybrid morphology consisting of layered nanosheets and flaky structures. Adsorption experiments were conducted to determine the optimal conditions for caffeine removal, with results showing the best adsorption efficiency at pH 7 and an adsorbent dosage of 0.1 g/L for the nanocomposite, achieving an 80.73% removal efficiency. The maximum adsorption capacity of the nanocomposite was 108.33 mg/g—significantly higher than for pure CB (55.05 mg/g) and PANI (57.71 mg/g). The adsorption process followed the pseudo-second-order kinetic model and the Langmuir isotherm, indicating a chemisorption mechanism and monolayer adsorption. Additionally, the nanocomposite demonstrated excellent reuse capacity, maintaining over 85% of its initial efficiency after multiple adsorption–desorption cycles, highlighting its potential for sustainable long-term use. This work demonstrates the potential of using waste-derived materials like cuttlefish bone as an effective support for PANI in the development of low-cost, sustainable adsorbents for pollutant remediation in pharmaceutical wastewater. Future studies will explore the adsorbent’s applicability for other contaminants and its potential antimicrobial properties.

1. Introduction

In recent years, it has been established that the emerging contaminants, pharmaceuticals, and personal care products present in waters are problematic compounds in terms of disposal. Pharmaceutical and personal care products (PPCPs) are classified as emerging pollutants because of their widespread use. Because these pollutants are found in surface waters and present health risks when ingested due to their high toxicity and bioaccumulation properties regarding biological tissues, impacting both human and animal health, they have garnered a lot of study interest [1]. Such substances are not removed completely by conventional methods of treatment; additionally, when bioaccumulated, they pose a potential risk to human health and aquatic animals. One of these substances is caffeine. The alkaloid caffeine (1,3,7−trimethylxanthine) is a member of the methylxanthine family. Because of its stimulating action on the central nervous system, which temporarily reduces drowsiness, it is commonly ingested via coffee, tea, chocolate products, and sodas, in which it is naturally found. Caffeine also affects the cardiovascular and pulmonary systems, making it a risk factor for people who have cardiovascular disease. Additionally, it can increase the effects of some medications and induce depression and hyperactivity [2]. The average amount of caffeine consumed globally is about 70 mg per person, but consumption varies by country. Around 5% of the caffeine that is consumed is eliminated through the urine and eventually finds its way into waterways through sewage systems, as well as via the disposal of caffeine-containing foods, drinks, and medications [3]. Caffeine may therefore be used to monitor the pollution level in a water source. Caffeine, for instance, has been found in natural waters in Germany and Italy at levels ranging from 0.6 to 1.056 and 80 to 265 ng/L, respectively [4].

A wide range of technologies have been developed to address the removal of these contaminants from water [5]. The major technologies include precipitation–coagulation, membrane separation, ion exchange, and adsorption. Adsorption techniques are widely used to remove several pollutants from water, especially those that are not easily biodegradable, and these methods can be used in small-scale household units [6]. Among the various approaches used to remediate emerging water pollutants, adsorption has gained significant attention due to its effectiveness and versatility. Techniques such as membrane separation [7], bioremediation [8], and advanced oxidative processes have been explored [9], but adsorption [9], particularly using novel adsorbents, offers distinct advantages in removing contaminants like caffeine from aqueous solutions [6,10,11].

Polyaniline (PANI) was selected for this study due to its unique properties that enhance the adsorption of contaminants like caffeine. As a conducting polymer, PANI exhibits excellent redox and electrical conductivity, which allow for the modulation of its surface charge, improving interactions with pollutants [12,13,14]. The presence of amine groups in PANI facilitates adsorption through hydrogen bonding, electrostatic interactions, and π−π stacking, making it highly effective for removing organic pollutants from aqueous solutions [13]. Moreover, PANI’s stability across a range of pH values and its ease of synthesis make it a cost-effective and versatile material [15]. By combining PANI with cuttlebone, we take advantage of both the adsorptive capacity of the natural material and the enhanced surface properties provided by PANI, resulting in an efficient and sustainable adsorbent for caffeine remediation. Furthermore, PANI is widely used in water treatment and as an antibacterial agent because it is effective against bacteria, fungi, and other microorganisms [16,17,18]. In addition, the interior cartilaginous shell found in octopus, squid, and cuttlefish is called the cuttlebone (CB). CB powder has long been used as a medication to treat various ear illnesses, prevent bleeding, and enhance kidney function [19]. It is a naturally occurring biomaterial that may be powdered from the cuttlefish’s chamber. Chitin and calcium carbonate (87.3–91.75%) represent the majority of the substances found in CB. Furthermore, it has trace amounts of copper, silicon, manganese, aluminum, titanium, and barium [20]. Biological cuttlefish bone (CB) has already received a significant amount of attention, with the aim of improving the sustainability of the material’s production by utilizing economical secondary raw materials to stop resource depletion [21]. Since CB is an inorganic material, it may be calcined to create calcium oxide, which is useful for many applications (such as biomaterials for pharmaceuticals and bone implants) [22]. These materials have potential uses in several practical domains, including energy and environmental applications [23]. Thus, CB has been modified with the pyrolysis technique and has the potential to produce activated carbon/CaO, which is useful as a potential biomaterial and adsorbent [24].

CB has shown potential for adsorbing other pollutants, including heavy metals and organic compounds. Several studies have reported its use as an adsorbent for dyes, heavy metals like lead and cadmium, and other organic contaminants, further emphasizing its versatility and environmental benefits [25,26]. By combining cuttlebone (CB) with polyaniline (PANI) in a homogeneous layer, we aimed to explore whether this nanocomposite material could serve as a cost-effective adsorbent. This approach not only improves the surface properties of the CB but also significantly enhances its adsorption capacity for a wide range of pollutants, including caffeine. The key strength of our work lies in the dual benefit of utilizing waste materials and enhancing their performance through functionalization, making the resulting adsorbent both sustainable and highly effective. An analysis of zeta potentials, Fourier-transform infrared (FTIR), X-ray diffraction (XRD), scanning electron microscopy with energy-dispersive X-ray (SEM−EDX), BET, and particle sizes were used to investigate the high adsorption capacity and its surface structure.

2. Materials and Methods

2.1. Chemicals

Aniline (C₆H₇N, 99.7%), ammonium persulfate (APS, (NH₄)₂S₂O₈, 99.7%), and hydrochloric acid (HCl, 38%) were purchased from Sigma-Aldrich (St. Louis, MO, USA) for use in this study. The CB used was a by-product of Sepia officinalis supplied by a local fish shop in Egypt. CB was obtained from the Governorate of Alexandria, whereas potassium hydroxide (KOH) (Oxford, India, 99%) and sodium hydroxide (NaOH, laboratory grade) were obtained from Egyptian Piochem (Giza) for laboratory chemical use without further purification. Hydrochloric acid (HCl, 38%) was obtained from CarloErba reagents (Memmingen, Germany) in Egypt and used as received, without further purification. Methanol with purity 99% was provided by ALPHA CHEMIKA−India (Mumdai).

2.2. Cuttlebone (CB) Preparation

Cuttlefish bones, collected as marine debris from Alexandria Governorate Beach in Alexandria, Egypt, were cleaned to remove odor and microbes. The bones were washed thoroughly and divided into small fragments. For the cleaning process, 50 g of CB was mixed with 500 mL of deionized water. The CB fragments were then boiled for five min, dried at 100 °C for 12 h, and pulverized into powder using a ball mill (ZK Corp Mini Ball Mill, Zhengzhou, China) with a cylinder size of 910 mm in diameter and 1120 mm in length, and a loading weight capacity of 0.2 tons. The milling parameters included a rotation speed of 35 RPM and a milling time of 1 h [27]. After milling, the CB powder underwent pyrolysis for 5 h in a tube furnace under a nitrogen atmosphere at 900 °C, (CIT17/Prog Delhi, India). Once the calcination process was complete, the furnace was allowed to cool to room temperature. The specimens were then stored in a desiccator prior to analytical evaluation [27,28,29].

2.3. Preparation of Polyaniline (PANI)

PANI was prepared using the method of chemical oxidation polymerization. In a typical preparation, 1.6 mL of aniline was added to 50 mL of deionized water and stirred for 30 min to form a homogeneous solution. Next, 50 mL of an acidic solution (prepared by dissolving 36.0 mM ammonium persulfate and 36.0 mM HCl in deionized water) was added dropwise to the aniline solution under constant stirring. The polymerization reaction was carried out for approximately 2 h at 0 °C. After polymerization, methanol was used to remove oligomers from the PANI solution. All reagents were carefully measured to ensure reproducibility, and the experimental conditions, such as temperature and concentrations, were maintained throughout the process.

2.4. Polyaniline (PANI)/Cuttlebone (CB) Nanocomposite Preparation

PANI/CB was fabricated by adding 1 g of PANI and 1 g of pyrolyzed CB to methanol. The mixture was stirred, and then an equal volume of distilled water (25 mL of water for every 25 mL of methanol) was added. After that, the mixture was dried at 60 °C for 12 h to allow interaction between the two materials, as illustrated in Scheme 1 (the mass ratio of PANI/CB was 1:1). The solid product of PANI/CB was then collected by vacuum filtration, washed multiple times, and dried at 60 °C (Scheme 1).

2.5. Characterization of the Prepared Materials

XRD was carried out on a PANalytical (Empyrean, Malvern Panalytical, Malvern, UK) utilizing Cu Ka radiation (wavelength 0.154 nm) at an accelerating voltage of 40 kV and a current of 35 mA. FTIR spectra were obtained using (Bruker-Vertex 70, KBr pellet technique, Karlsruhe, Germany) (serial number 1341) to determine the molecular vibration in the chemical bonds between 400 and 4000 cm⁻¹ wavenumbers. Scanning electron microscopy (SEM) images were acquired using Gemini Zeiss-Sigma 500 VP (Jena, Germany) to study the morphologies of the generated materials, while elemental analysis was performed using energy-dispersive X-ray (EDX) spectroscopy. which was carried out using a Quanta FEG250 instrument from Germany (Bremen), to assess the elemental composition of the nanocomposite N₂ adsorption. An automated surface analyzer from Micrometrics in the USA (TriStar II 3020, Norcross, GA, USA) was used for this analysis.

2.6. The Adsorption Studies

To perform this experiment, we applied the batch operating system at an ambient temperature using a standard stock aqueous solution of caffeine of concentration 500 µg/mL, which allowed for the preparation of a series of diluted concentrations to obtain an ideal calibration curve (5–200 µg/mL). The removal of caffeine from solution using the PANI/CB nanocomposite adsorbent was carried out by mixing 100 µg/mL of caffeine with 0.2 g of PANI/CB in a 50 mL Falcon tube using bi-distilled water. The mixture was incubated for 300 min at room temperature, with the solution maintained at a pH of 7. Afterward, the PANI/CB adsorbent was separated from the solid particles through centrifugation. The residual concentration of caffeine was measured using a double-beam UV–visible spectrophotometer (UV−2600 UV−Vis Spectrophotometer, Shimadzu) at wavelength 272 nm; the (RE%) and the adsorption capacity (q_e) were calculated using Equations (1) and (2), respectively.

$$RE\%={\displaystyle \frac{C_i-C_e}{C_i}}\times 100$$

(1)

$$q_e={\displaystyle \frac{(C_i-C_e)}{m}}\times V$$

(2)

where RE is the removal efficiency (%) and C_i and C_e are the caffeine’s initial and equilibrium concentrations (mg/L), respectively. q_e is the equilibrium capacity of the adsorption (mg/g), V represents the volume of the solution (L), and m corresponds to the adsorbent mass (g).

The adsorption behavior of caffeine on cuttlefish bone (CB), polyaniline (PANI), and their nanocomposite (PANI/CB) was studied using eight non-linear equilibrium isotherm models. These included three two-parameter models, Langmuir, Freundlich, and Dubinin−Radushkevich (D−R), which describe monolayer adsorption, heterogeneous adsorption, and surface heterogeneity, respectively. Four three-parameter models, including Langmuir−Freundlich, Sips, Redlich−Peterson, and Toth, were used to account for mixed adsorption behaviors and varying surface energies. A four-parameter Baudu model was applied for multilayer adsorption systems. For the isotherm analysis, caffeine solutions with varying initial concentrations were prepared and added to fixed amounts of CB, PANI, or PANI/CB. The mixtures were stirred until equilibrium was reached, after which the equilibrium concentration of caffeine was measured. The amount of caffeine adsorbed (q_e) was plotted against the equilibrium concentration (C_e) for each adsorbent. The experimental data were fitted to the isotherm models using non-linear regression. Model performance was evaluated based on the correlation coefficient (R²) and root mean square error (RMSE) to determine the best-fitting model.

2.7. Regeneration of Adsorbent

To evaluate the reusability of the PANI/CB nanocomposite, a series of adsorption and desorption cycles were conducted. After each adsorption cycle, the adsorbent surface was thoroughly washed three times with distilled water to remove any residual pollutants and then dried at room temperature. Fresh solutions containing DMF, N−heptanol, ethanediol, acetone, ethyl acetate, di−ethyl ether, and ethanol (each with an adsorbate concentration of 10 mg/L) were then introduced for each new cycle to regenerate the adsorbent. This procedure ensured that the adsorbent was evaluated with a new solution for each cycle to assess its reusability under consistent experimental conditions. The regeneration efficiency was evaluated by measuring the caffeine recovery from the adsorbent surface after each desorption phase, and the process was repeated for five consecutive cycles. The performance of the PANI/CB nanocomposite was determined using Equation (1).

3. Results and Discussion

3.1. Characterization of Prepared Materials

3.1.1. Scanning Electron Microscopy (SEM)

PANI (PANI), CB (CB), and nanocomposite (PANI/CB) surface morphology were examined using scanning electron microscopy (SEM). Figure 1A,B show the morphology of PANI after polymerization, where the formation of structures is evident. However, agglomeration has obscured the initial porous structure and stacked nanosheets [30]. On the other hand, Figure 1C,D show SEM imaging of the structure of CB. The particle size of CB was heterogeneous in nature and varied in size from fine (around 5–10 µm) to large flaky particles (approximately 20–50 µm) [31]. Figure 1E,F show the SEM imaging of the PANI/CB nanocomposite structure, where the two phases are easily identifiable: the first is the layered nanosheet form of PANI, and the second is CB, which contains flakes. The prepared material has the potential to aggregate or cluster together, resulting in the formation of larger particles or structures [32]. This aggregation can reduce the exposed surface area available in the nanocomposite, resulting in a decrease in the BET surface area from 92.64 m²/g for the CB to 31.58 m²/g for the PANI/CB. The observed decrease in surface area from 92.64 m²/g for CB to 31.58 m²/g for the PANI/CB nanocomposite is not solely due to PANI’s lower surface area but can also be attributed to the aggregation of CB particles, which reduces the available surface area for adsorption; see Table 1 for the BET analysis results. As a result, the surface area of the nanocomposite (PANI/CB) is reduced compared to pure CB, but still greater than that of PANI alone.

The chemical composition of PANI, CB, and PANI/CB nanocomposites was identified using the EDX spectrum. According to the results, PANI was composed of carbon (C), nitrogen (N), and oxygen (O) (Figure 2A). The CB powder spectrum (Figure 2B) showed peaks for carbon (C), oxygen (O), magnesium (Mg), calcium (Ca), phosphorus (P), zirconium (Zr), iodine (I), and titanium (Ti). While the presence of these elements, particularly Zr, I, and Ti, is not typically expected in cuttlebone, it could be attributed to impurities in the cuttlebone source or artifacts of the EDX analysis. The PANI/CB nanocomposite spectrum (Figure 2C) showed peaks for C, O, K, Ca, N, and Mg. The presence of characteristic peaks for carbon (C), nitrogen (N), and oxygen (O) in all three spectra confirms the presence of PANI. Additionally, the presence of calcium (Ca) and phosphorus (P) peaks in the PANI/CB and CB spectra indicates the presence of cuttlebone. While quantitative analysis of carbon content based on peak intensities can be challenging due to factors like sample thickness and instrument settings, the relative peak intensities suggest a higher carbon content in the PANI/CB nanocomposite compared to pure PANI, consistent with the incorporation of carbon-rich cuttlebone. The EDX results, in conjunction with the SEM and BET analysis, provide strong evidence for the successful integration of PANI onto the surface of cuttlebone, resulting in a nanocomposite material with enhanced adsorption properties.

3.1.2. Fourier-Transform Infrared (FTIR)

Figure 3 depicts the Fourier-transform infrared (FTIR) spectra of PANI, CB, and PANI/CB. The FT−IR spectrum of CB is displayed in Figure 3A; raw CB shows weak OH and NH group absorption peaks in the range of 3651–3445 cm⁻¹, attributed to the stretching vibration of water molecules [31,33]. The NH group’s absorption is due to the interaction of water molecules or other surface-bound species on the CB, not due to nitrogen in the CB itself. Strong peaks at 1128, 885, and 702 cm⁻¹ reveal the presence of aragonite structure carbonate groups n1, n2 (out-of-plane bend), and n4 (doublet C−O in-plane bend), respectively [34]. A broad peak around 1441 cm⁻¹ confirms the n3−asymmetric stretch of the aragonite carbonate structure [35]. The absorption peak of PANI at 3468 cm⁻¹ is assigned to the stretching vibration of the amine group, and bands at 2340 cm⁻¹ reveal hydrogen bonding. The peak at 1577 cm⁻¹ is assigned to the stretching vibration of the aromatic ring. The peak at 1486 cm⁻¹ corresponds to the N−H stretching vibration, while the peak at 1318 cm⁻¹ is assigned to the C−N stretching vibration [36]. The distinctive peaks detected in the FT−IR spectra of the PANI/CB nanocomposite provide vital information about the PANI conformation within the CB channels and potential interactions between CB and PANI. The changes observed in the FTIR peaks are indicative of interactions involving both bonding (such as hydrogen bonding) and possible intermolecular interactions between CB and PANI. The FT−IR spectrum of the PANI/CB nanocomposite (Figure 3C) shows characteristic bands of both PANI and CB, confirming the presence of both components in the nanocomposite. Notable changes in the peak positions and intensities are observed, which are caused by the interaction between the constituents of the nanocomposite. This suggests that CB interacts with the polymer matrix through both hydrogen bonding and other intermolecular interactions [37].

3.1.3. X-Ray Diffraction (XRD)

The XRD patterns of CB, PANI, and their nanocomposite (PANI/CB) are shown in Figure 4. The diffractogram of CB exhibits several peaks, with the most significant peak observed at around 30°, which is typically associated with the characteristic diffraction pattern of aragonite (calcium carbonate polymorph). However, it is important to note that the XRD spectrum for CB shows multiple reflections that are consistent with different polymorphs of calcium carbonate. The most prominent peaks correspond to the (104), (110), (113), and (020) planes, characteristic of aragonite, calcite, and other calcium carbonate forms. The intensity of the peaks observed in the diffractogram of CB is relatively low, which may be attributed to factors such as the material’s crystallinity or sample preparation conditions. In particular, the lower intensity can be related to the nature of the sample and the experimental setup, rather than the smaller particle size. To explain this result, it is more likely that the observed lower intensity is due to the inherent characteristics of the sample and experimental conditions, such as the material’s crystallinity, sample preparation, or scanning parameters. Accurate peak indexing is essential for properly identifying and representing the crystalline phases present in the sample [33,38]. The observed broad peaks and lower definition are typical for natural materials that may not crystallize well under standard conditions. For reference, these polymorphs are aligned with the JCPDS cards for aragonite (JCPDS #47−1743) and calcite (JCPDS #47−1741). For PANI, the XRD pattern aligns well with the data from the JCPDS card (#53−1891) and shows a characteristic peak at 2θ = 6.5°, which can be attributed to the long-range ordering of the polymer chains and the overall polymeric structure [39,40]. This peak is indicative of a more ordered structure in the polymeric chains of PANI, although it is possible that a semi-crystalline structure exists, as the lower signal intensity could also be influenced by experimental conditions such as scan rate or sample amount [41]. In the PANI/CB nanocomposite, the XRD diffractogram exhibits a broad peak centered at 2θ = 27.2° corresponding to the (003) plane, which is characteristic of the polymeric nature of PANI and confirms the successful incorporation of PANI into the nanocomposite material. Although PANI was synthesized first and CB was added afterward, the interaction between the PANI backbone chain and CB during the composite formation process can still influence the crystallization process. The presence of CB may act as a nucleating agent, potentially facilitating crystallization and contributing to the small increase in crystallite size observed in the nanocomposite. Additionally, the increased intensity of XRD peaks in the nanocomposite compared to PANI indicates a strong interaction between PANI and CB, which may contribute to the enhanced structural stability and functionality of the nanocomposite material [42].

3.1.4. The N₂ Adsorption−Desorption Isotherms

The N₂ adsorption−desorption isotherms over the surface of the samples are depicted in Figure 5A. Figure 5B also shows the pore size distribution (PSD) curves. Table 1 summarizes the surface parameters acquired from the N₂ adsorption−desorption isotherms. The obtained isotherms of PANI (PANI), CB (CB), and the nanocomposite (PANI/CB) have IV isotherms with distinct hysteresis loops, which are typical of mesoporous structures. The three samples exhibit H₃ hysteresis loops, which can be described as the aggregation of plate-like particles with split-shaped pores [43]. The inflection points at P/P₀ = 0.9 in a sorption−desorption isotherms analysis signify the transition from monolayer formation to multilayer formation. Sorption−desorption isotherm analysis typically indicates the completion of the first monolayer of adsorbate on the surface of the adsorbent. At low relative pressures, the adsorbate molecules are strongly attracted to the surface of the adsorbent. They form a monolayer, and the adsorption rate is rapid. As the relative pressure approaches 0.9, the surface becomes increasingly covered with the first monolayer, indicating a transition in the adsorption behavior. At this point, the surface is nearing saturation, and the rate of adsorption begins to change. This signifies that most of the available surface area has been occupied by the adsorbed molecules, and additional gas molecules are beginning to fill in the multilayer. The adsorption rate slows down because there are fewer available sites for adsorption [44,45]. The CB has a greater surface area (92.64 m²/g) than the PANI/CB (31.58 m²/g). The decrease in the surface area of the nanocomposite sample related to aggregation/agglomeration significantly reduces the interfacial/interphase and tensile properties of nanocomposites via decreasing the specific surface area. This aggregation can reduce the exposed surface area available in the nanocomposite, resulting in a decrease in the BET surface area from 92.64 m²/g for the CB to 31.58 m²/g for PANI/CB. As a result of aggregation, the porosity of the nanocomposite sample is enhanced, which is reflected in the pore geometry data (Table 1) [46,47].

Table 1. Surface parameters for the samples.

Sample	BET Surface Area (m2/g)	Total Volume in Pores (cm3/g)	Total Area in Pores (m2/g)
CB	92.64	0.16	56.98
PANI	13.14	0.02	7.26
PANI−CB	31.58	0.06	20.47

Table 1. Surface parameters for the samples.

Sample	BET Surface Area (m2/g)	Total Volume in Pores (cm3/g)	Total Area in Pores (m2/g)
CB	92.64	0.16	56.98
PANI	13.14	0.02	7.26
PANI−CB	31.58	0.06	20.47

3.1.5. Surface Texture and 3D Characterization

Surface morphology has been extensively studied because it can display vital properties, including deformations and heterogeneities that can affect the material’s application. Mountain Map^® 9.0 software was used to analyze the topographic SEM image. The surface profile analysis of PANI, CB, and PANI/CB revealed high roughness “peaks” and “valleys.” Figure 6 exhibits surface 3D SEM micrographs (left) and the Abbott−Firestone curve and the depth histogram of the samples (right). The height (or depth) distribution is characterized by a histogram, which implies the likelihood (frequency) of points being at a certain height (or depth). The Abbott−Firestone curve is marked in red, with the vertical axis graduated in depths and the horizontal axis in percentages of the overall population. The three samples have a distinct height distribution. The surface texture directions of the samples were analyzed using Cartesian graphs (Figure 7), as well as the related values of surface texture parameters, which may be associated with the inhomogeneity of the surface [42].

Table 2. Textural characteristics of three samples.

Parameters	Values of Parameters
	PANI	CB	PANI/CB
Total roughness (Rt) (nm)	52.12	0.63	170.00
Fractal dimension (Df)	1.170	1.418	1.109
Slope	0.996	0.998	0.998
Roughness skewness (Rsk)	0.218	−0.587	0.303
Roughness kurtosis (Rku)	1.881	2.916	2.608

shows the surface parameters in terms of roughness (R_a), roughness skewness (Rsk), roughness kurtosis (Rku), and fractal dimension (D_f). The isotropy percentages of PANI, CB, and PANI/CB are 88.06, 83.12, and 95.98%, respectively. This is in line with the aspect ratio of the texture (Str) value of 0.676, 0.830, and 0.928, respectively, which indicates that the surface texture of the three samples is isotropic. If Str is close to the unit, the surface is isotropic, and if Str is close to 0, the surface is anisotropic [48]. The kurtosis and skewness parameters are processed using the square root of the surface height distribution (RMS). Roughness kurtosis (Rku) is a measure of the “sharpness” of a surface and the unpredictability of profile heights. Spiky surfaces have Rku > 3, bumpy surfaces have Rku < 3, and absolutely random surfaces have Rku = 3.

Accordingly, the values of Rku are 1.88, 2.916, and 2.608 for PANI, CB, and PANI/CB, respectively, which indicates that the three samples have a rough surface. A roughness skewness (Rsk) value reflects the symmetry of the surface; a negative value implies a predominance of valleys, while a positive value indicates a “peaky” surface. PANI and PANI/CB have a positive Rsk, indicating that the surfaces of PANI and PANI/CB are peaky. Meanwhile, the CB has a negative value, indicating that the surface of the CB is a valley surface. The fractional dimension (D_f) quantifies the complexity of a fractal sample. The correlation coefficient (R²) of the linear fit equality between the enclosed area and the scale of analysis in this study was close to 1, indicating that the data were excellently fitted by linear functions. The values of D_f are 1.170, 1.418, and 1.109, as listed in Table 2. The decrease in D_f indicates the formation of more regular structures [49].

3.2. Significant Parameters Affect the Adsorption of Caffeine on PANI, CB, and PANI/CB

3.2.1. Effect of pH

The adsorption of caffeine onto cuttlefish bone (CB), polyaniline (PANI), and their nanocomposite (PANI/CB) is significantly influenced by pH, which is related to the point-of-zero charge (PZC) of each adsorbent and the pKa values of caffeine. Caffeine has pKa values of 0.8, 10.4, and 14.0, and remains predominantly neutral around pH 7, making this pH range critical for adsorption studies. The PZC values for the adsorbents are as follows: cuttlefish bone (PZC = 8.2), polyaniline (PZC = 5.5), and their nanocomposite (PZC = 7.0). As shown in Figure 8, the maximum caffeine adsorption occurs near pH 7, where all three adsorbents exhibit near-neutral surface charges. This minimizes electrostatic repulsion and promotes non-electrostatic interactions, such as van der Waals forces and hydrogen bonding. At pH levels above or below 7, the adsorbent surfaces become charged either positively or negatively, which increases electrostatic repulsion and disrupts the adsorption process. Therefore, the adsorption assays were performed at pH 7 to maximize caffeine removal, as both caffeine and the adsorbents are in their neutral states under these conditions [50,51,52].

3.2.2. Effect of Dose of Adsorbent

The impact of adsorbent dose on caffeine removal efficiency was investigated by testing doses ranging from 0.025 g to 0.2 g per 50 mL of solution (Figure 9) [53,54]. In our experiments, a caffeine concentration of 100 mg/L was used, which is a standard concentration for similar adsorption studies. The adsorption conditions included a solution pH of 7, a contact time of 120 min, and room temperature. The results, illustrated in Figure 9, demonstrate a clear trend: as the dose of the adsorbent increases, the removal efficiency of caffeine also rises. This increase continues until a dose of 0.1 g for the nanocomposite and 0.2 g for both polyaniline and cuttlefish bone are reached, where each adsorbent achieves its peak efficiency. Beyond these doses, the increase in removal efficiency becomes less pronounced, with a tendency towards leveling off. This indicates that while higher doses continue to increase the removal efficiency, the improvement becomes smaller as more active sites on the adsorbent are utilized. At higher doses, the available active sites are increasingly occupied, leading to diminishing returns in removal efficiency. This finding underscores the importance of optimizing the dose of adsorbent to balance between maximizing removal efficiency and avoiding an unnecessary excess of material. Carrying out the experiment with these optimal doses ensures effective utilization of the adsorbent while avoiding wastage and maintaining cost efficiency.

3.2.3. Adsorption Isotherm

The adsorption behavior of caffeine on cuttlefish bone (CB), polyaniline (PANI), and their nanocomposite (PANI/CB) was analyzed using eight non-linear equilibrium isotherms models (Figure 10). These models include three two-parameter isotherms: Langmuir, Freundlich, and Dubinin–Radushkevich (D−R). Additionally, four three-parameter models were used: Langmuir−Freundlich, Sips, Redlich−Peterson, and Toth. Lastly, one four-parameter model, Baudu, was applied. The performance and suitability of these models in fitting the experimental data were assessed and are summarized in

Table 3. The non-linear adsorption isotherm models using three adsorbents.

Isotherm Models	Expression	Adjustable Model Parameters	Cuttle Fish Bone (CB)	R2	Poly Aniline (PANI)	R2	Nanocomposite (PANI/CB)	R2
Two-parameter isotherm
Langmuir	$q_e={\displaystyle \frac{q_{max} K_L{C}_e}{1+K_L{C}_e}}$	qmax	55.05	0.99	57.71	0.99	96.99	0.99
		KL	0.0067		0.006		0.003
Freundlich	$q_e=K_f{C}_e^{{\displaystyle \frac{1}{nf}}}$	Kf	2.02	0.98	0.508	0.98	0.61	0.99
		1/nf	0.48		2.01		1.35
Dubinin–Radushkevich (D−R)	$q_e=q_{m\exp (-K_{ad {\pounds }^2})}$	qm	45.43	0.97	46.34	0.97	68.78	0.96
		Kad	0.012		0.01		0.019
Three-parameter isotherm
Langmuir−Freundlich	$q_e={\displaystyle \frac{q_{max}{\left(K_{LF C_e}\right)}_{{\beta }_{LF}}}{1+{\left(K_{LF}{C}_e\right)}_{{\beta }_{LF}}}}$	qmax	50.01	0.99	62.74	0.99	108.33	0.99
		KLF	0.008		0.005		0.015
		βLF	1.14		0.91		0.82
Sips	$q_e={\displaystyle \frac{q_{max K_{s{C}_e{n}_s}}}{1+K_{s C_{e n_s}}}}$	qmax	50.01	0.99	62.70	0.99	103.15	0.99
		KS	0.004		0.008		0.015
		ns	1.14		0.92		0.65
Redlich−Peterson	$q_e={\displaystyle \frac{K_{RP C_e}}{1+a_{RP C_e^{\beta RP}}}}$	KRP	0.30	0.99	0.41	0.99	0.35	0.99
		aRP	0.0008		0.011		0.008
		β	1.29		0.93		0.87
Toth	$q_e={\displaystyle \frac{K_{e C_e}}{\left[1+{(K}_{TCe{)}^{nT}}\right]\frac{1}{nT}}}$	Ke	0.39	0.99	0.39	0.99	0.31	0.99
		KT	0.0008		0.011		0.004
		nT	1.29		0.93		0.94
Four-parameters isotherm
Baudu	$q_{e }={\displaystyle \frac{q_{max b_{0C_e^{1+x+y}}}}{{1+b}_{0 C_e^{1+x}}}}$	qmax	50.01	0.99	57.70	0.99	93.61	0.99
		bo	0.004		0.006		0.0029
		X	0.00001		0.00001		0.0001
		Y	0.14		0.00001		0.037

. The cuttlefish bone and polyaniline materials exhibited adsorption patterns consistent with the Langmuir and Freundlich models, suggesting that the adsorption processes are governed by a combination of physical and chemical interactions, with a tendency for both monolayer adsorption (Langmuir) and heterogeneous surface adsorption (Freundlich). These findings align with previous studies, which have reported similar adsorption behaviors for caffeine on various adsorbents, such as activated carbon and bio-based materials [55,56,57]. The PANI/CB nanocomposite demonstrated a more complex adsorption behavior, as indicated by the better fit to the Langmuir−Freundlich, Sips, Redlich−Peterson, and Toth models, which suggests a more versatile adsorption mechanism. These models account for a combination of heterogeneous adsorption sites and the cooperative adsorption effect, which is typical for nanocomposites with multiple functional groups or varying surface characteristics. The higher adsorption capacity of the nanocomposite can be attributed to the synergistic effects of both PANI and CB, which may provide more active sites and enhance the overall adsorption processes.

The statistical analysis revealed that throughout the whole concentration range, the two-parameter isotherm Sips had the best correlation with the experimental data. This was because the statistical analysis’s determined error functions had the lowest values and the highest coefficient of determination (R²), which was almost equal to one. In order to predict isotherm modeling in sets of eight adsorption systems, HYBRID and the chi-squared test (χ²) were thought to be universal indicators that offered the best fit to the experimental data. The results obtained from the remaining statistical criteria varied, and their nature might depend on the number of experimental points, model parameters, and pressure range. The statistical error validity data for the isotherm models are shown in

Table 4. Summary of the determined error functions for the nonlinear adsorption isotherm models for caffeine using CB.

Parameter	Langmuir	Freundlich	Dubinin−Radushkevich	Langmuir−Freundlich	Redlich	Sips	Toth	Baudu
SSE/ERRSQ	12.49	51.15	38.30	10.92	8.02	10.92	8.02	10.92
χ2	1.65	2.95	6.64	2.06	1.71	2.06	1.71	2.06
Adjusted R2	0.98	0.91	0.93	0.98	0.99	0.98	0.99	0.98
MAE	1.20	2.41	1.92	0.96	0.68	0.96	0.68	0.96
MAPE/ARE	13.16	16.79	26.80	13.36	11.54	13.36	11.54	13.36
RMSE	1.34	2.70	2.34	1.25	1.07	1.25	1.07	1.25
RMSE_2	1.58	3.20	2.77	1.48	1.27	1.48	1.27	1.48
NRMSE	0.13	0.27	0.23	0.13	0.11	0.13	0.11	0.13
HYBRID	18.42	23.50	37.52	18.70	16.15	18.70	16.15	18.70
HYBRID_2	32.97	58.99	132.78	41.17	34.13	41.17	34.14	41.17
HYBRID_3	1.65	2.95	6.64	2.06	1.71	2.06	1.71	2.06
MPSD	25.59	23.14	51.80	30.49	28.19	30.49	28.19	30.49
MPSD_2	0.33	0.27	1.34	0.46	0.40	0.46	0.40	0.46
SAE/EABS	8.39	16.88	13.45	6.73	4.75	6.73	4.75	6.73
RMS	21.63	19.56	43.78	25.77	23.83	25.77	23.83	25.77
NSD	0.22	0.20	0.44	0.26	0.24	0.26	0.24	0.26
ARE_2	4.68	3.83	19.17	6.64	5.68	6.64	5.68	6.64
ARE_3	8.18	7.39	16.55	9.74	9.01	9.74	9.01	9.74

,

Table 5. Summary of the determined error functions for the nonlinear adsorption isotherm models for caffeine using PANI.

Parameter	Langmuir	Freundlich	Dubinin−Radushkevich	Langmuir−Freundlich	Redlich	Sips	Toth	Baudu
SSE/ERRSQ	6.98	30.64	64.23	6.17	6.62	6.17	6.62	6.98
χ2	1.34	2.33	8.89	1.06	1.25	1.06	1.25	1.34
Adjusted R2	0.99	0.95	0.90	0.99	0.99	0.99	0.99	0.99
MAE	0.78	1.87	2.77	0.71	0.71	0.71	0.71	0.78
MAPE/ARE	11.04	15.34	33.25	9.95	10.45	9.95	10.44	11.04
RMSE	1.00	2.09	3.03	0.94	0.97	0.94	0.97	1.00
RMSE_2	1.18	2.48	3.58	1.11	1.15	1.11	1.15	1.18
NRMSE	0.10	0.21	0.30	0.09	0.10	0.09	0.10	0.10
HYBRID	15.45	21.47	46.55	13.93	14.63	13.93	14.62	15.45
HYBRID_2	26.85	46.51	177.83	21.14	24.97	21.14	24.95	26.85
HYBRID_3	1.34	2.33	8.89	1.06	1.25	1.06	1.25	1.34
MPSD	24.80	23.13	57.98	21.28	23.65	21.28	23.63	24.80
MPSD_2	0.31	0.27	1.68	0.23	0.28	0.23	0.28	0.31
SAE/EABS	5.44	13.12	19.36	4.98	4.99	4.98	4.99	5.44
RMS	20.96	19.55	49.00	17.98	19.98	17.98	19.97	20.96
NSD	0.21	0.20	0.49	0.18	0.20	0.18	0.20	0.21
ARE_2	4.39	3.82	24.01	3.23	3.99	3.23	3.99	4.39
ARE_3	7.92	7.39	18.52	6.80	7.55	6.80	7.55	7.92

and

Table 6. Summary of the determined error functions for the nonlinear adsorption isotherm models for caffeine using PANI−CB.

Parameter	Langmuir	Freundlich	Dubinin−Radushkevich	Langmuir−Freundlich	Redlich	Sips	Toth	Baudu
SSE/ERRSQ	23.07	9.48	197.01	10.87	19.29	8.38	21.40	26.71
χ2	3.23	0.53	19.74	1.42	2.84	0.57	3.13	3.75
Adjusted R2	0.98	0.99	0.82	0.99	0.98	0.99	0.98	0.97
MAE	1.60	0.87	4.90	1.14	1.37	0.86	1.46	1.66
MAPE/ARE	19.13	5.93	46.88	12.98	17.85	6.84	18.61	20.27
RMSE	1.82	1.16	5.31	1.25	1.66	1.09	1.75	1.95
RMSE_2	2.15	1.38	6.28	1.47	1.96	1.29	2.07	2.31
NRMSE	0.15	0.10	0.45	0.11	0.14	0.09	0.15	0.17
HYBRID	26.78	8.30	65.63	18.17	24.99	9.58	26.05	28.38
HYBRID_2	64.56	10.63	394.79	28.30	56.81	11.40	62.56	75.05
HYBRID_3	3.23	0.53	19.74	1.42	2.84	0.57	3.13	3.75
MPSD	33.53	9.33	72.45	22.03	31.75	10.99	33.18	36.01
MPSD_2	0.56	0.04	2.62	0.24	0.50	0.06	0.55	0.65
SAE/EABS	11.19	6.06	34.28	8.00	9.60	6.01	10.23	11.60
RMS	28.34	7.88	61.23	18.62	26.84	9.29	28.04	30.43
NSD	0.28	0.08	0.61	0.19	0.27	0.09	0.28	0.30
ARE_2	8.03	0.62	37.50	3.47	7.20	0.86	7.86	9.26
ARE_3	10.71	2.98	23.14	7.04	10.14	3.51	10.60	11.50

. These isotherm models were validated using five statistical techniques. It has been found that the better the model, and the lower the values of these statistical tools, the closer the agreement between the experimental quantity adsorbed (qe, exp) and the calculated quantity adsorbed (qe, cal). There was a mention of the coefficient of regression (R²) to support and compare the best model. The isotherm models’ ability to describe the sorption process is improved by higher R² values, closer qe, exp and qe, cal values, and lower statistical error values, which fitted to Sips for caffeine adsorption (Table 5 and Table 6).

3.2.4. Adsorption Kinetics

The kinetic study of caffeine adsorption onto cuttlefish bone, polyaniline, and their nanocomposite revealed that the nanocomposite exhibited superior adsorption performance (Figure 11). All three adsorbents followed pseudo-second-order kinetics, indicating that chemisorption was the rate-limiting step [58]. The nanocomposite achieved the highest adsorption capacity (51.49 mg/g) and the highest correlation coefficient (R² = 0.98), suggesting a more favorable adsorption process. This is consistent with studies in the literature [50,59], where pseudo-second-order kinetics have frequently been observed for caffeine adsorption on various adsorbents, including composites, where the adsorption process is primarily governed by chemical interactions. The performance and suitability of these models in fitting the experimental data were assessed and are summarized in

Table 7. The kinetic models’ parameters.

Kinetic Models	Equation	Parameters	Cuttlefish Bone	Polyaniline	Nanocomposite
Pseudo-first-order	qt = qe (1 − e−k1t)	K1	0.015	0.044	0.043
		Qe	38.93	40.24	49.31
		R2	0.97	0.98	0.99
Pseudo-second-order	$qe={\displaystyle \frac{q{e}^{2}k2t}{1+qek2t}}$	K2	0.0006	0.001	0.001
		Qe	41.70	42.15	51.49
		R2	0.97	0.98	0.98
Mixed 1,2 order	$qt=qe \left({\displaystyle \frac{1-\exp \left(-kt\right)}{1+f2\exp (-kt)}}\right)$	K	0.0006	0.02	0.035
		Qe	39.58	40.63	49.44
		f2	0.73	0.68	0.26
		R2	0.97	0.98	0.99
Avrami	qt = qe [1 − exp(−kavt)nav]	Qe	38.93	40.25	49.31
		Kav	0.25	0.43	0.42
		nav	0.06	0.10	0.10
		R2	0.97	0.98	0.99
Intraparticle diffusion	qt = Kip√t + Cip	Kip	0.95	0.86	1.08
		Cip	14.02	19.05	22.56
		R2	0.70	0.54	0.53

. Additionally, the mixed 1,2 order model also indicated the nanocomposite’s superiority, with a higher F2 value (0.73) compared to the other adsorbents. The mixed 1,2 order model suggests that the adsorption process involves both surface adsorption and diffusion, which may occur simultaneously, reflecting a more complex interaction mechanism. This mixed kinetic model has been reported in the literature for caffeine adsorption systems where the process is governed by both chemical interactions at the surface and diffusion mechanisms within the adsorbent structure [60,61]. The intraparticle diffusion model, however, revealed that diffusion was not the sole rate-limiting step for the nanocomposite, suggesting that additional factors, such as chemical bonding or electrostatic interactions, also play a significant role in the adsorption process [62]. These findings highlight the multifaceted nature of the adsorption process and explain the superior performance of the nanocomposite in caffeine removal [6,59].

3.2.5. Effect of Regeneration and Reusability

The reusability of an adsorbent is a critical factor in designing an eco-friendly and sustainable adsorption process for contaminants. The reusability performance of the PANI/CB nanocomposite was evaluated through a series of sequential cycles of adsorption and desorption experiments (Figure 12A). Among the solvents tested, N−heptanol demonstrated the highest regeneration efficiency, successfully recovering 95% of the caffeine attached to the adsorbent surface. However, it is worth noting that acetone also exhibited a significant regeneration efficiency, suggesting that it could be a viable alternative to N−heptanol [50]. Acetone’s effectiveness can be attributed to its ability to disrupt the intermolecular forces between the adsorbate and the adsorbent surface, likely through hydrogen bonding and polarity interactions [63]. This results in a high desorption efficiency, facilitating the removal of caffeine from the adsorbent surface [64]. Following the regeneration process, a desorption phase was initiated using an optimum solvent to release the adsorbed material, enabling the recovery of the adsorbate. This process was repeated for five cycles, meticulously monitoring changes in adsorption capacity. The magnitude of caffeine removed after each adsorption cycle was expressed by the removal percentage. Figure 12B illustrates the efficiency of caffeine removal following each regeneration cycle. By the conclusion of the fifth consecutive adsorption cycle, the caffeine removal rates decreased from 85% to 75% for the PANI/CB nanocomposite. The decrease in adsorption efficiencies may be linked to the reduction in surface chemistry groups following each adsorption cycle. As the cycles progressed, the adsorbent surface likely underwent changes, such as the loss of functional groups or a reduction in the availability of active sites, leading to weaker π−π interactions and diminished electron transfer, which are crucial for the adsorption process [65,66]. This decrease in performance aligns with findings in the literature, where similar adsorbents show reduced capacity over successive cycles due to surface degradation and site saturation [67,68,69].

4. Conclusions

In conclusion, this research highlights the significant potential of using cuttlebone waste as a sustainable, eco-friendly resource for developing adsorbent materials. The study demonstrates the effectiveness of cuttlebone as both an adsorbent and a supporting phase for polyaniline (PANI). SEM analysis of PANI, cuttlebone (CB), and the PANI/CB nanocomposite revealed distinct surface morphologies. PANI exhibited agglomerations that obscured its original porous structure and stacked nanosheets, while CB particles were heterogeneous in size, ranging from fine to large, flaky particles. The BET surface area showed a decrease from 92.64 m²/g for CB to 31.58 m²/g for PANI/CB. Under optimal conditions (room temperature, 0.2 g/L adsorbent dose, and pH 7), the caffeine removal efficiency was 62.54%, 66.98%, and 80.73% for CB, PANI, and PANI/CB, respectively. Adsorption kinetics were best described by the Mixed 1,2 order and Avrami models. Langmuir isotherms indicated maximum caffeine adsorption capacities of 50.01, 62.74, and 108.33 mg/g for CB, PANI, and PANI/CB, respectively. This innovative approach not only offers a solution to waste management issues but also contributes to the development of novel, efficient adsorbents. The promising results open new avenues for research and applications, encouraging the further exploration of natural and waste-derived materials for advanced environmental solutions.