The biopolymer, sodium alginate (SA) having a formula weight of 176.10 g/mg and molecular formula C₆H₈O₆, acrylic acid (AAc, 99 %), N,N′-methylene bis-acrylamide (MBA, 99%), ammonium persulfate (APS) (≥98%; 248614), and MWCNTs were procured from Sigma-Aldrich (South Africa). Acetone, sodium hydroxide (NaOH), nitric acid (HNO₃, 65%), hydrochloric acid (HCl) and MB were supplied by Merck (South Africa). All chemicals used in this study were of analytical grade and used without any further purification. Deionized (DI) water was used throughout the experiment. Typically, 1.0 g of MB powder was dissolved in 1000 ml volumetric flask in DI water, stock solutions were further prepared by diluting the as-prepared stock solution from batch experiments.

The FTIR spectra of SA, o-MWCNTs, SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC are shown in Figure 1A. It is noticeable that o-MWCNTs displayed a broad characteristic peak at 3,666 cm⁻¹ to 3,015 cm⁻¹, corresponding to the–OH groups. The sharp peaks at 2,922 cm⁻¹ and 2,841 cm⁻¹, attributed to the asymmetric stretching and symmetric vibrations of CH₂ group (Dyke and Tour, 2004). The characteristic bands corresponding to C-O and–COO groups at 1,035 cm⁻¹ and 1608 cm⁻¹, respectively. The spectrum of SA revealed a wide band between 3,100 and 3,620 cm⁻¹ related to the stretching vibration of the hydroxyl groups present in the SA polymer chain. The peak at 2,922 cm⁻¹ was assigned to the –CH, which stretching vibrations of the CH₂ (Makhado et al., 2017a; Makhado et al., 2017b). Two absorption peaks at about 1,608 and 1,407 cm⁻¹ were ascribed to the COO⁻ groups and the C-C bending vibration, respectively. The band at 1,034 cm⁻¹ emanated from the C-O-C stretching frequencies (Hu et al., 2018; Wattie et al., 2018). The characteristic peak at 802 cm⁻¹ relates to the Na-O of the SA (Thakur et al., 2016; Olad et al., 2018). After graft co-polymerization of AAc onto SA, an increase in the peak intensity of SA was observed at 802 cm⁻¹ which was attributed to the β-glycosidic linkage between the guluronic units and a very broad absorption peak appeared at 3,660–2,500 cm⁻¹ due to carboxylic acid OH stretch. Furthermore, the shift in the position of the characteristic peaks of SA shifted from 1,680 to 1,700 cm⁻¹ which is possibly due to the strong interaction with AAc and the stretching of the C-O-C bond shifted from 1,023 to 1,121 cm⁻¹ which confirms the SA/p(AAc) hydrogel formation (Thakur et al., 2016). In the spectrum of SA/p(AAc)/o-MWCNTs HNC, the characteristic peaks emanating from SA/p(AAc) hydrogel were all observed with a reduction in the intensity after the formation of hydrogel nanocomposite, these results suggest that the internal hydrogen bonding occurred between the o-MWCNTs and SA/p(AAc) hydrogel.

The XRD patterns of pure MWCNTs and o-MWCNTs are illustrated in Supplementary Figure S1B. Both samples showed three diffraction peaks at 2θ = 25.8° relating to the (002) plane of the graphitic carbon (JCPDS-ICDD No. 751621), 2θ = 43.1° corresponding to the (001) and 2θ = 53.3° belonging to the (004) reflections of the carbon atoms. The diffraction patterns of samples exhibited the same trend without any shift in the 2θ peak positions, which indicate that the microstructure of MWCNTs was not destroyed during oxidizing. Moreover, the intensity of diffraction peaks of the o-MWCNTs spectrum was slightly higher than that of pure MWCNTs, confirming that pure MWCNTs were oxidized (Supplementary Figure S1B). These observations were in good agreement with the results obtained from Raman studies (Supplementary Figure S1A). XRD spectra of pure SA, SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC are displayed in Figure 1B. In the case of pure SA, 2θ diffraction peaks at 13.8°, 21.5°, and 39.1° can be attributable to the reflection of (110) plane from polyguluronate unit, (200) plane from polymannuronate and the latter from amorphous halo, respectively. After graft copolymerization of AAc onto SA backbone, the peaks emanating from SA emerged and formed a broad diffraction peak at 2θ = 18.9°, without any clear diffraction peaks confirming the amorphous nature of SA/p(AAc) hydrogel. The disappearance of two diffraction peaks suggested successful grafting of AAc onto SA backbone. Several studies showed that grafting of AAc onto biopolymer results in the interference of the backbone diffraction peaks (Thakur et al., 2016; Verma et al., 2020). Subsequent SA/p(AAc)/o-MWCNTs HNC formation, the appearance of a wide broad peak with higher intensity was observed at 2θ = 18.9° and a diffraction peak at 2θ = 43.2° corresponding to (100) reflection, which may be due to a short-range order in o-MWCNTs layers. The XRD results are in good agreement with the FTIR results, confirming the formation of SA/p(AAc) hydrogel as well as SA/p(AAc)/o-MWCNTs HNC.

The TGA of o-MWCNTs, SA, SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC is shown in Figure 1C. The thermal characteristics of the o-MWCNTs showed high stability over the studied range 25–600°C. It was observed that SA exhibits two degradation steps. In the first step, weight loss of about 10% occurred between 50 and 180°C which may be attributed to dehydration of SA and volatilization of water. The major weight loss of around 40% occurred over the range 250–300°C, which was mainly due to the SA backbone decomposition. The thermal characteristics of SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC demonstrated similar degradation profiles with curves disintegrated in three steps. The first degradation of 5% occurred in the temperature range of 90–210°C due to the loss of water residual. In the temperature region of 250–350°C, there was a second decomposition stage of about 20% due to degradation of SA and residue carbide form (Soares et al., 2004). In the third degradation profile, about 50% weight reduction occurred in the temperature range of 330–500°C due to the breaking of its carboxyl from acrylic acid groups and degradation of the residual copolymer. Overall, the correlation analysis demonstrated that SA/p(AAc)/o-MWCNTs HNC had improved thermal stability as compared to SA/p(AAc) hydrogel. The presence of o-MWCNTs in the hydrogel matrix enhanced the thermal property of the hydrogel.

Structural morphologies of SA, SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC was subjected to SEM analysis. Figures 2A,B depicts low and high magnification of SA. Fragment-like loose surface was observed for the pure SA, which is indicative of semi-crystalline and amorphous structure. After the formation of the three-dimensional polymer network, the surface morphology presented a relatively continuous coarse surface which is a consequence of the presence of AA (Figures 2C,D). Figures 2E,F illustrated the low and high magnification of SA/p(AAc)/o-MWCNTs HNC. It was observed that subsequent hydrogel nanocomposite formation, the surface morphology of became utterly different from the continuous coarse surface to rough, irregular and porous structure. The rough, irregular and porous structure results from interactions between o-MWCNTs and polymeric hydrogel matrices.

To get information on the homogeneity of o-MWCNTs on the SA/p(AAc) hydrogel matrix, TEM analysis was performed and the results are shown in Figures 3A,B. From the micrographs, it is seen that o-MWCNTs were mostly uniformly dispersed on the SA/p(AAc) hydrogel matrix to form SA/p(AAc)/o-MWCNTs HNC. The o-MWCNTs were compatible with the polymeric matrices due to availability of hydrophilic functional groups, which promote strong H bonding interactions.

The viscoelastic property of polymeric hydrogel is an important factor which affects their applications. The storage modulus, G′ (elastic response) and loss modulus, G″ (viscous behavior) of the SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC were measured as a function of frequency using DMA. Figure 4A shows the G′ of SA/p(AAc) hydrogel and its nanocomposite in the 0.01–100 Hz frequency range and similar trends were noticed. It can be seen that both SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC exhibited elastic character at low frequencies. The SA/p(AAc)/o-MWCNTs HNC revealed lower elastic response than SA/p(AAc) hydrogel. This result suggests that the incorporation of o-MWCNTs in the polymeric hydrogel matrices weakened the degree of cross-linking density and caused the decrease in G′. Figure 4B displays the viscous behavior of SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC as a function of frequency. The G″ of SA/p(AAc) hydrogel was slightly higher than that of SA/p(AAc)/o-MWCNTs HNC in the frequency range of 0.01–30 Hz, cross over was noticed at 30 Hz. As the frequency of the system was increased above the characteristic frequency of the segments, the dissipative component increased. It can be observed from Figures 4A,B that the storage modulus was greater than the loss modulus throughout the studied frequency range, which was indicative of elastic character.

The water uptake of SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC was investigated in the pH range 3.0–9.0 (Figure 4C). The swelling studies results demonstrated that both adsorbents had pH-dependent behavior. Moreover, both adsorbents showed similar trends for the water uptake throughout the studied pH range. In acidic conditions, both adsorbents displayed the lowest swelling degree. It was observed that the increase in the solution pH increased the swelling capacity of SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC. The maximum swelling percentage of 1,669.7 and 2,268.4% was observed at pH 8.0 for SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC, respectively. In the alkaline pH, the network structure of both adsorbents exhibited extensive swelling in the aqueous medium due to the presence of hydrophilic groups, i.e., -COOH, which renders a negative charge density that participates in the electrostatic interactions with the water molecules. The o-MWCNTs incorporated hydrogel nanocomposite exhibited greater swelling capacities than the SA/p(AAc) hydrogel. The increment in the swelling degree of SA/p(AAc)/o-MWCNTs HNC can be ascribed to the addition of hydrophilic o-MWCNTs in the hydrophilic polymeric matrices (Makhado et al., 2018b). The swelling degree of SA/p(AAc)/o-MWCNTs HNC was comparatively high, which is owing to the low gel strength. It is well proven that the improvement in the gel strength is inversely proportional to the swelling degree (Thakur et al., 2016). Therefore, the swelling studies results support the mechanical studies results.

The influence of salts (0.1 M of NaCl and CaCl₂) on the swelling capacity of SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC was investigated and the results are shown in Figure 4D. It was observed that both adsorbents exhibited high swelling capacity in the absence of salts. The maximum swelling capacity of 136.0 and 150.1 g/g for SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC was recorded for the solution of NaCl, respectively. In the solution of CaCl₂, the highest water uptake was found to be 124.8 and 149.7 g/g for SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC. These observations demonstrate that salts reduced the water molecule adsorption efficiency of both adsorbents. These reductions may be attributed to the competition between water molecules and the surface of the adsorbent. The pH value at the point of zero charges (pHpzc) on the SA/p(AAc) hydrogel surface was found to be 2.2 and the surface become negatively charged with increasing pH values. On the other hand, SA/p(AAc)/o-MWCNTs HNC demonstrated negatively charged surface in the studied pH range and the negative charge on the surface of SA/p(AAc)/o-MWCNTs HNC increased with increasing pH values (Figure 4E). These results may be explained by dissociation of the hydroxyl groups that imparts the negative charge to the MWCNTs surface. The negatively charged surface promote the adsorption of positively charged adsorbate such as MB. A similar behaviour was reported in the literature (Makhado et al., 2018b).

Various factors influencing the adsorption process such as solution pH, adsorption dose, contact time and initial dye concentration in the solution must be optimized for application in real-world situations.

The influence of solution pH plays a significant role in the adsorption process. The solution pH governs the interaction between the ions present in the reaction system and the adsorbent surface. Figure 5A showed that the adsorption capacity increased with the increasing solution pH and subsequently the equilibrium was reached at pH 6.0. FTIR results confirmed that SA/p(AAc)/o-MWCNTs HNC has carboxylic groups which could be attributed to the high MB affinity at pH above 5.0. In an acidic medium, the carboxylic groups on the SA/p(AAc)/o-MWCNTs HNC surface were protonated limiting MB uptake due to columbic repulsion. With an increase in the solution pH, electrostatic interaction between negatively charged SA/p(AAc)/o-MWCNTs HNC surface and cationic MB dye molecules enhanced the adsorption capacity. In a basic medium, the high adsorption capacity was obtained due to ionization of carboxylic groups at higher pH, favoring intermolecular interaction between adsorbate and adsorbent. Therefore, the optimum solution pH used for the adsorption of MB in this study was 6.0. Similar observations were noticed with SA/p(AAc) hydrogel.

It is conspicuous in Figure 5B that the adsorption capacity was decreased with the increasing the amount of SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC in the adsorption process. The adsorption capacity of SA/p(AAc)/o-MWCNTs HNC decreased from 1,540 to 1,350 mg/g as the adsorbent dose was increased from 0.33 to 1.67 g/l at a fixed solution pH, contact time and equilibrium concentration. The high adsorption capacity was achieved at low adsorbent dose due to complete occupation of the active sites. At high adsorbent loading, the adsorption capacity decreased due to less occupation of adsorption active sites and agglomeration of the particles in the reaction mixture. SA/p(AAc)/o-MWCNTs HNC showed a high adsorption capacity as compared to the neat SA/p(AAc) hydrogel throughout the studied range.

The influence of contact time on the adsorption capacity of MB by SA/p(AAc) hydrogel and SA/p(AAc)/o-MWCNTs HNC was evaluated in the range of 10–50 min at solution pH of 6.0, 0.33 g/l adsorbent dose and 600 mg/l at three different temperatures. As depicted in Figure 5C, the adsorption capacity of MB was rapid and then, attained equilibrium within 40 min. The adsorption kinetic models were applied to investigate the adsorption mechanism of adsorbent and adsorbate. To better understand the adsorption process, the experimental data in Supplementary Figure S2C were evaluated using two common kinetic models namely pseudo-first-order and pseudo-second-order. As shown in Figures 6A,B and Table 1, the pseudo-second-order displays the correlation coefficient (R²) of 0.999, which is higher than that of the pseudo-first-order model. Moreover, the calculated RMSE values favored the pseudo-second-order model as a probable kinetic model. These results and kinetic theory, suggest that MB adsorption onto SA/p(AAc)/o-MWCNTs HNC was governed by chemisorption and physicochemical interactions (Makhado et al., 2019b).

The effect of the of initial concentration MB dye on the adsorption capacity of the SA/p(AAc)/o-MWCNTs HNC was evaluated in the range 530–1,350 mg/L of under optimum conditions (pH 6.0, 50 min contact time and a dose of 0.33 g/L) at three different temperatures, and the results are presented in Supplementary Figure S2D. It was observed that the adsorption capacity of SA/p(AAc)/o-MWCNTs HNC increased with increasing temperature, which implied that the adsorption process was endothermic in nature. The increase in adsorption capacity with increasing initial concentration demonstrated that the mass gradient functions as the driving force for the adsorption process. Furthermore, an increase in the concentration of MB dye solution at fixed SA/p(AAc)/o-MWCNTs HNC amount raised the mass transfer process.

To understand the relationship between the initial MB concentration and SA/p(AAc)/o-MWCNTs HNC the experimental adsorption data were fitted directly to three isotherm models namely Langmuir, Freundlich and Temkin models Figures 7A–C, respectively. The related isotherm parameters were calculated and tabulated in Table 2. The best-fitting adsorption isotherm model was selected based on the correlation coefficient value. The Langmuir isotherm model was the best fit for describing the adsorption process, which assumes that the adsorption process was monolayer coverage and the adsorption site on the SA/p(AAc)/o-MWCNTs HNC surface was homogeneous. The maximum adsorption capacities determined by Langmuir isotherm were found to be 1,596.0, 1,673.9, and 1,838.9 mg/g at 298.15, 308.15, and 318.15 K, respectively. The applicability of the Langmuir isotherm model was supported by the values of R_L, which were less than 1 for the sorption of MB dye. Moreover, the RMSE values for Langmuir isotherm were found to be least among other studied other isotherm models, confirming that the Langmuir isotherm model is an excellent one. A brief list of the reported adsorbents, used for the removal of MB from aqueous solution as well as the obtained maximum adsorption capacity, including the prepared adsorbents in this study, is presented in Table 3 (Mittal et al., 2015; Sun et al., 2015; Mittal and Ray, 2016; Dai et al., 2017; Mahdavinia et al., 2017; Makhado et al., 2018a; Makhado et al., 2018b; Hassan et al., 2019; Ribeiro et al., 2019; Makhado et al., 2020). These results suggest that the SA/p(AAc)/o-MWCNTs HNC may be applied for the adsorption of MB from the aqueous phase. Moreover, SA/p(AAc)/o-MWCNTs HNC exhibited improved adsorption capacity as compared to unmodified hydrogel as shown in Figure 5D. SA/p(AAc)/o-MWCNTs HNC showed better maximum adsorption capacity than the one reported for SA/p(AAc) hydrogel (Makhado et al., 2020).

The spontaneity of the MB dye adsorption using the prepared SA/p(AAc)/o-MWCNTs HNC at a temperature ranging from 298.15 to 318.15 K was evaluated to obtain thermodynamic parameters. Thermodynamics parameters such as enthalpy change (∆H°), entropy change (∆S°), and Gibbs free change (∆G°) were investigated via Vant Hoff’ equation, which is expressed as shown in Eqs 16,17: Δ G ° = Δ H ° − T Δ S ° , (16) ln ( q e C e ) = Δ H ° RT − Δ S °   R , (17) where R is a universal gas constant (8.314 J/mol K), T is the temperature (K) and (qe/ce) is equilibrium constant. The values of ∆H° and ∆S° were calculated from the slope and intercept of the linear plot between In (q_e/c_e) and 1/T shown in Figure 8A, the values of ∆G^° were obtained from Eq. 16 and all thermodynamic parameters are summarized in Supplementary Table S1. The positive value of ∆S° and ∆H° indicated that the adsorption process was disordering and randomness at the solid-liquid interface on the adsorbent. Improved sorption efficiency, as well as the endothermic process, was validated by the positive value of ∆H°. The negative ∆G° values showed the spontaneous nature of the sorption process (Dogan et al., 2009). Moreover, the decrease in values of ∆G° with increasing temperature suggests that the sorption is favorable at high temperatures.

The desorption of MB dye molecules adsorbed on the surface of SA/p(AAc)/o-MWCNTs HNC was conducted using HCl solution. The applicability of the hydrogel nanocomposite was investigated in six adsorption-desorption cycles and the results are shown in Figure 8B. The obtained results showed that the SA/p(AAc)/o-MWCNTs HNC maintained high adsorption percentage even after the sixth successive cycles under the same conditions. In conclusion, SA/p(AAc)/o-MWCNTs HNC can be used repeatedly for adsorption of MB dye without a significant decrease in MB dye affinity. This adsorbent holds great potential for remediation of MB wastewater.

Figure 8C shows the plausible mechanism for the adsorption of MB dye onto SA/p(AAc)/o-MWCNTs HNC. Both hydrogel and filler (o-MWCNTs) consist of carboxyl groups on the surface. Therefore, the formation of the SA/p(AAc)/o-MWCNTs HNC was achieved via strong hydrogen bond between hydrogel and o-MWCNTs. The adsorption studies (effect of solution pH) showed that high adsorption of MB onto the SA/p(AAc)/o-MWCNTs HNC surface was achieved at pH range 5–9. This result suggests strong electrostatic attraction between the adsorbent and the MB dye molecules. Abundant of carboxyl groups on the surface of SA/p(AAc)/o-MWCNTs HNC dissociated into carboxylate anions, which influenced attraction of positively charged MB dye molecules. The π-π stacking interaction between MB onto o-MWCNTs may also be responsible for the high sorption capacity of SA/p(AAc)/o-MWCNTs HNC.