Synthesis and Application of Novel Poly(Acrylic Acid–Maleic Acid)/Graphene Oxide Nanocomposite Hydrogel [P(AA-MA)/GO] for Efficient Adsorptive Removal of Azur A Dye from Aqueous Solutions

INTRODUCTION
Water contamination has become one of the most important environmental problems facing modern society and has led to substantial scientific efforts in developing effective cleanup solutions [1, 2, 3]. Synthetic dyes are a particularly hazardous group of these due to their vast employment in textile, paper, leather and pharmaceutical manufacturing industries [3]. Azur A is one of the cationic thiazine dyes, commonly utilised for biological staining and textile colouration and has received much attention due to its resistant character and reported deleterious effects on living systems [4]. Some methods studied for dye removal from the contaminated streams are advanced oxidation processes, membrane separation, biological degradation and adsorption-based methods [2, 5, 6]. Among these alternatives, adsorption has emerged as a highly attractive choice, owing to its operational simplicity, cost-effectiveness, and exceptional efficacy in eliminating a wide variety of contaminants, all without producing secondary waste streams [6].
Hydrogel materials based on crosslinked polymer networks have shown great potential in water purification applications [1, 7]. Polyacrylic acid (PAA) has a large number of carboxylic acid groups, it is highly hydrophilic and has good pH-responsive swelling behaviour, hence it is especially promising for adsorption applications [8]. Maleic acid, with its two carboxyl groups, is a great comonomer to boost the crosslinking density and to raise the concentration of anionic sites [9]. Graphene oxide (GO) is a two-dimensional carbon-based nanomaterial adorned with oxygen-containing functional groups which has exhibited a remarkable adsorption capacity towards a variety of organic contaminants [10, 11]. The combination of poly(acrylic acid–maleic acid) hydrogel with graphene oxide nanosheets offers a new technique to synergistically utilise the advantages of both components. To the best of our knowledge, this nanocomposite system has not been reported before for the removal of Azur A dye. The present study focuses on the synthesis of P(AA-MA)/GO nanocomposite hydrogel via free radical polymerisation and its full characterisation as a possible adsorbent for the removal of Azur A dye, along with extensive research on swelling and reusability.

MATERIALS AND METHODS
Materials and Characterization Techniques
All chemicals employed were of analytical grade. Acrylic acid (AA), maleic acid (MA), N,N′-methylenebisacrylamide (MBA), potassium persulfate (KPS), graphite powder, concentrated sulfuric acid, hydrochloric acid, potassium permanganate, hydrogen peroxide, sodium hydroxide, and sodium chloride were obtained from Sigma-Aldrich, Merck, Fluka, BDH, and Scharlau. Azur A dye (C.I. Basic Blue 15, C₁₄H₁₄ClN₃S, M = 287.8 g/mol) was obtained from Himdia (India). Deionized water was used throughout.
FTIR spectroscopy was performed on a Shimadzu 8400S (400–4000 cm⁻¹). XRD analysis used a Shimadzu XRD-6000 (Cu-Kα, λ = 1.5406 Å, 40 kV, 30 mA, 2θ = 10–80°). FE-SEM employed a TESCAN MIRA3 (20 kV). AFM measurements used an ARA-Research ARA-AFM in contact mode. Nitrogen adsorption–desorption isotherms at 77 K were measured using a Quantachrome NOVA 2200e. TGA was performed on a Perkin Elmer TGA4000 (25–700 °C, 10 °C/min, N₂). UV-Vis spectrophotometry used a Shimadzu UV-1800 at λmax = 647 nm. A calibration curve (A = 0.1016 × C − 0.0109; R² = 0.9978) was constructed for Azur A quantification.

Synthesis of GO and Hydrogel
Graphene oxide was prepared by the modified Hummers method [12, 13]. Briefly, 1.0 g graphite + 1.0 g NaNO₃ + 46 mL conc. H₂SO₄ were stirred in an ice bath; 6.0 g KMnO₄ was added slowly. After 30 min at 35 °C, the mixture was diluted with 92 mL water (reaching ~98 °C), then 280 mL water and 10 mL of 30% H₂O₂ were added, producing a brilliant yellow suspension. The product was washed (10% HCl, then water), dialyzed for 3 days, and dried at 45 °C.
The crosslinked hydrogel was prepared via free-radical copolymerization. Maleic acid (0.02 mol, 2.321 g) and MBA (0.005–0.02 g) were dissolved in 5 mL water. Acrylic acid (0.08 mol, 5.764 g) and KPS (0.015 g) were added dropwise under N₂. Polymerization proceeded at 60 °C for 2 h. The nanocomposite hydrogel was prepared analogously by predispersing GO (0.02–0.16 g per 0.1 g nanocomposite) in the aqueous medium prior to polymerization. The product was washed and dried at 40 °C.

Adsorption Experiments
Batch experiments were conducted by adding adsorbent (0.001–0.1 g) to 10 mL Azur A solution (100–700 mg/L). Effects of pH (3–9), dose, GO content, time (1–400 min), concentration, temperature (5–30 °C), and ionic strength (NaCl, KCl, CaCl₂) were investigated. Residual dye was quantified at 647 nm. Removal efficiency (R%) and equilibrium adsorption capacity (Qe) were calculated as R(%) = (C₀ − Ce)/C₀ × 100 and Qe = (C₀ − Ce)·V / m.

Swelling and Reusability Studies
For swelling: 0.05 g of the dried nanocomposite was immersed in 50 mL of buffered aqueous solutions (pH 3–10); at predetermined time intervals, samples were blotted and weighed. Swelling ratio: SR(%) = (Ws − Wd)/Wd × 100. For reusability: five consecutive adsorption–desorption cycles were carried out under optimum conditions. After each adsorption cycle, the dye-loaded nanocomposite was regenerated with 0.1 M HCl/ethanol (50:50 v/v) for 60 min, washed, and re-used.

RESULTS AND DISCUSSION
FTIR Analysis
FTIR spectra of GO, pristine P(AA-MA), and P(AA-MA)/GO before and after Azur A adsorption are shown in Fig. 2. GO exhibits O–H stretching at 3440 cm⁻¹, C=O at 1720 cm⁻¹, C=C at 1620 cm⁻¹, C–OH at 1220 cm⁻¹, and C–O–C at 1060 cm⁻¹ [10, 11, 13]. The pristine hydrogel displays a broad O–H/N–H band at 3500–3000 cm⁻¹, aliphatic C–H at 2930 cm⁻¹, free carboxylic C=O at 1720 cm⁻¹, amide C=O at 1650 cm⁻¹, and carboxylate vibrations at 1560 and 1400 cm⁻¹ [8, 9]. Upon GO incorporation, the carbonyl peak shifts from 1720 to 1690 cm⁻¹ indicating hydrogen bonding between the polymer carboxyl groups and GO oxygen functionalities [14, 15]. After Azur A adsorption, new peaks at 1580 and 1495 cm⁻¹ (aromatic C=C and C–N of the thiazine ring) confirm successful adsorption, with shifts in carboxylate peaks indicating electrostatic interactions [4, 10].

XRD Analysis
XRD patterns (Fig. 3) show that pristine graphite exhibits a sharp peak at 2θ = 26.50° (d = 3.36 Å) [10, 11]. After oxidation, this peak disappears and a new characteristic peak appears at 2θ = 11.60° (d = 7.60 Å), confirming successful oxidation [10, 13]. The pristine P(AA-MA) hydrogel shows a broad amorphous hump at 21.13° (d = 4.19 Å) [8]. In the nanocomposite, this broad peak shifts slightly to 20.20° (d = 4.39 Å) and the GO characteristic peak at 11.60° disappears entirely, indicating successful exfoliation and uniform dispersion of GO nanosheets within the polymer matrix [14-17].

FE-SEM Analysis
FE-SEM images (Fig. 4) show that graphite has a smooth, compact layered structure [10], whereas GO exhibits wrinkled, crumpled sheets [10, 11]. The pristine hydrogel displays a porous, sponge-like network [8]. The P(AA-MA)/GO nanocomposite shows a notably rougher surface with well-dispersed GO sheets embedded in the polymer matrix [9, 14]. After Azur A adsorption, the surface appears smoother and more uniform, with reduced porosity due to dye coverage [4, 15, 17].

AFM Analysis
AFM 3D topographical images (Fig. 5) confirm the contrast in surface morphology between the pristine hydrogel and the nanocomposite. The pristine P(AA-MA) hydrogel surface is relatively smooth, while the P(AA-MA)/GO nanocomposite exhibits a markedly rougher and more heterogeneous topography due to the protrusion of GO nanosheets at the surface [14, 15, 18].

BET Surface Area and Pore Structure
Nitrogen adsorption–desorption isotherms and pore size distributions are shown in Fig. 6. The BET surface area of GO is 318.06 m²/g with pore volume 0.270 cm³/g and pore diameter 1.937 nm. The pristine P(AA-MA) hydrogel has a microporous structure (SBET = 13.23 m²/g, Vp = 0.0026 cm³/g, dp = 1.850 nm). Notably, the P(AA-MA)/GO nanocomposite exhibits significantly enhanced textural properties: SBET = 25.72 m²/g (94.5% increase), with the average pore diameter expanding to 6.060 nm — shifting the material from a microporous to a mesoporous regime [14, 15, 17, 18]. Textural properties are summarized in Table 1.

TGA Analysis
TGA thermograms (Fig. 7) reveal three distinct decomposition stages for all samples. GO shows poor thermal stability (~91.7% total loss): 3.0% (40–90 °C, water desorption), 69.7% (90–248 °C, oxygen functional group decomposition), and 19.0% (248–593 °C, carbon skeleton degradation). The pristine hydrogel shows three stages: 2.39% (40–200 °C), 53.03% (200–343 °C), and 36.50% (343–600 °C). The nanocomposite shows similar behavior (2.49%, 56.07%, 38.62%), with a slight reduction in thermal stability attributed to GO’s high thermal conductivity [8, 9, 14, 15]. Detailed TGA data are presented in Table 2.

Effect of Solution pH and pHpzc
The effect of pH on Azur A adsorption (Fig. 8) shows a progressive increase in Qe from 145.44 mg/g at pH 3 to 149.39 mg/g at pH 9, with corresponding removal efficiency rising from 88.79% to >95%. The point of zero charge was found to be about pHpzc = 4.5 by the pH-drift method (Fig. 9). At pH < 4.5, the surface is positively charged due to the presence of protonated −COOH groups which do not allow the attraction of the cationic Azur A. At pH > 4.5, the deprotonation leads to −COO− sites and the electrostatic attraction increases dramatically [4, 10]. The optimum operating state was pH 7, where the system exhibited a compromise of excellent adsorption efficacy, near neutral circumstances and practical relevance of real wastewater treatment.

Swelling Behavior of the P(AA-MA)/GO Nanocomposite
Equilibrium Swelling and Effect of pH
The equilibrium swelling ratio in the pH range 3–10 (Fig. 10) show a very dramatic pH dependence, increasing from 97.5% at pH 3 to 1648% at pH 10. Such a pronounced pH sensitivity is typical for polyelectrolyte hydrogels containing ionisable carboxylic acid groups [7, 8]. At low pH the –COOH groups are largely protonated and there is little electrostatic repulsion and little swelling [8]. At higher pH, progressive deprotonation generates –COO⁻ groups, leading to electrostatic repulsion, osmotic pressure gradient, and extensive water uptake [7, 9, 19].

Swelling Kinetics
Time-dependent swelling at pH 7 (Fig. 11) shows a rapid initial uptake during the first 60 minutes, reaching ~78% of equilibrium, followed by gradual approach to equilibrium at approximately 240 minutes [20, 21]. This biphasic behavior reflects the steep concentration gradient at early times followed by progressive filling of the polymer network until osmotic and elastic forces equilibrate [22].

Effect of Adsorbent Dosage
Increasing the adsorbent dose from 0.001 to 0.1 g (Fig. 12) caused removal efficiency to rise from 97.60% to 98.71%, while Qe decreased dramatically from 2927.96 mg/g to 29.61 mg/g due to dilution of the fixed dye mass over more sites [8, 10, 11]. A dose of 0.04 g was selected as the practical optimum, balancing efficiency (98.62%) and capacity (73.96 mg/g).

Effect of Graphene Oxide Content
The influence of GO loading on the adsorption capacity of the nanocomposite was demonstrated in Fig. 13, where a change in GO content from 0.02 to 0.16 g per 0.1 g nanocomposite was observed with the maximum Qe value being obtained as 29.74 mg/g at 0.08 g GO loading, and then a progressive reduction was observed at higher GO content. The first improvement can be attributed to the larger surface area, the oxygen functional groups and the π–π stacking sites [11, 14, 15]. At excessive loadings, GO aggregation and increased crosslinking reduce porosity and adsorption performance [8, 10, 11].

Effect of Contact Time and Kinetic Studies
Adsorption capacity increased rapidly during the first 60 minutes from 140.2 mg/g (1 min) to 147.55 mg/g (60 min), reaching equilibrium at ~120 min (Fig. 14a). The pseudo-first-order model gave a poor fit (R² = 0.6038) with calculated Qe = 3.21 mg/g, far from the experimental value (Fig. 14b). The pseudo-second-order model gave excellent agreement (R² = 1.000, Qe = 149.25 mg/g) with k₂ = 0.0374 g/mg·min and initial rate h = 833.33 mg/g·min (Fig. 14c), confirming chemisorption as the rate-controlling mechanism [23]. Kinetic parameters are summarized in Table 3.

Effect of Initial Concentration and Temperature
Adsorption isotherms at 5, 10, 20, and 30 °C with C₀ from 100 to 700 mg/L are shown in Fig. 15a. At 30 °C, Qe rose from 1.98 mg/g (100 mg/L) to 99.81 mg/g (700 mg/L). Temperature exerted a positive effect on Qe across all concentrations, indicating an endothermic process. The van’t Hoff plot (Fig. 15b, ln K vs. 1/T) is linear with R² = 0.9341, yielding ΔH° = +37.46 kJ/mol and ΔS° = +163.36 J/mol·K. ΔG° values were −7.74, −8.78, −10.93, and −11.69 kJ/mol at 5, 10, 20, and 30 °C, respectively (Table 4), confirming spontaneous and increasingly favorable adsorption at higher temperatures [11, 17].

Effect of Ionic Strength
Adding NaCl, KCl, or CaCl₂ (0–0.012 g) to the dye solution (Fig. 16) caused Qe to decrease from 148.19 mg/g (no salt) to 136.30, 140.69, and 142.95 mg/g, respectively. The reduction order NaCl > KCl > CaCl₂ reflects competition of monovalent cations with Azur A for the negatively charged sites and compression of the electrical double layer, while CaCl₂ exhibits weaker inhibition due to its larger hydrated radius and possible bridging effects between carboxylate groups [8, 10, 11, 15]. Overall losses (<8%) demonstrate robust performance in saline media.

Adsorption Isotherm Modeling
Equilibrium data at 25 °C were fitted to Langmuir, Freundlich and Temkin models (Fig. 17, Table 5). Contrary to popular assumptions, the Langmuir model gave a poor fit (R 2 = 0.6987) with an unphysical low Qmax of 0.844 mg/g . This shows that the underlying assumption of monolayer adsorption on a homogeneous surface is not applicable to this nanocomposite. The Freundlich model fitted very well (R² = 0.9021) with KF = 45.44 (mg/g)(L/mg)^(1/n) and 1/n = 0.503 suggesting favourable adsorption on heterogeneous surface. The best fit (R2 = 0.9782) was obtained by the Temkin model with BT = 3.44 J/mol and KT = 0.001 L/mg which indicates the existence of significant adsorbent-adsorbate interactions where the heat of adsorption reduces linearly with surface coverage [24]. The dominance of Temkin and Freundlich fits over Langmuir is consistent with the heterogeneous structure of the nanocomposite with different binding-energy sites (carboxylate groups, GO π-domains, hydroxyl/epoxy groups) rather than a uniform monolayer coverage.

Reusability of the Nanocomposite
Reusability is a key parameter for practical application. Over five consecutive adsorption–desorption cycles using 0.1 M HCl/ethanol (50:50, v/v) as the regenerating medium (Fig. 18), the nanocomposite retained 82.34% of its original removal efficiency after the fifth cycle (decreasing from 98.87% in cycle 1). Corresponding desorption efficiencies ranged from 94.6% (cycle 1) to 80.9% (cycle 5). The gradual decline is attributed to incomplete desorption, minor structural changes from repeated acid–base cycling, and possible restacking of GO nanosheets [6, 7, 10, 11, 25]. The retention of >82% capacity over five cycles, combined with simple synthesis, high capacity, and broad pH tolerance, demonstrates the nanocomposite’s potential for practical wastewater treatment applications. A comparison with other adsorbents reported in the literature is presented in Table 6.

CONCLUSION
A novel P(AA-MA)/GO nanocomposite hydrogel was successfully synthesized via free-radical copolymerization and applied for the removal of Azur A dye from aqueous solutions. Comprehensive characterization (FTIR, XRD, FE-SEM, AFM, BET, TGA) confirmed successful GO integration with enhanced surface area (25.72 m²/g) and a mesoporous structure (6.06 nm). Optimal adsorption (98.62% removal, 148.24 mg/g) was achieved at pH 7, 0.04 g dose, 120 min contact, and 25 °C. The nanocomposite exhibits pronounced pH-responsive swelling (97.5–1648%), pHpzc ≈ 4.5, and excellent reusability (>82% efficiency after 5 cycles). Kinetics follow the pseudo-second-order model (R² = 1.000), confirming chemisorption. The Temkin isotherm provides the best fit (R² = 0.9782) followed by Freundlich (R² = 0.9021), while Langmuir is a poor fit (R² = 0.6987) — indicating that adsorption proceeds on a heterogeneous surface with significant adsorbent–adsorbate interactions rather than uniform monolayer coverage. Thermodynamic analysis confirms a spontaneous (ΔG° = −7.74 to −11.69 kJ/mol), endothermic (ΔH° = +37.46 kJ/mol), entropy-driven (ΔS° = +163.36 J/mol·K) process. The proposed mechanism involves electrostatic attraction, hydrogen bonding, π–π stacking, pore filling, and hydrophobic interactions. These attributes position P(AA-MA)/GO as a promising adsorbent for dye-contaminated wastewater treatment.

CONFLICT OF INTEREST
The authors declare that there is no conflict of interests regarding the publication of this manuscript.