Graphene oxide incorporated cellulose acetate beads for efficient removal of methylene blue dye; isotherms, kinetic, mechanism and co-existing ions studies

In this investigation, new porous adsorbent beads were formulated via the incorporation of graphene oxide (GO) into cellulose acetate beads (CA) for the adsorptive removal of methylene blue (MB) dye. The experimental results signified that the adsorption of MB dye increased with the increase in the GO ratio from 10 to 25%. In addition, the adsorption process obeyed PSO kinetic model and Langmuir isotherm model with a maximum adsorption capacity reaching 369.85 mg/g. More importantly, it was proposed that the adsorption mechanism of MB dye onto GO@CA proceeded via electrostatic interactions, H-bonding, van der Waals forces, n-π and π -π interactions. Besides, the fabricated beads exhibited an excellent ability to recycle and reuse after five successive cycles. In addition, there was a high selectivity of GO@CA beads towards MB molecules in the presence of co-existing cations such as Fe2+, Zn2+, Cu2+ and Ni2+.

Graphene oxide (GO) is a carbonaceous material that is easily prepared via oxidation of graphite and possesses remarkable merits such as extremely high specific surface areas (theoretical S BET = 2620 m 2 /g), a good mechanical characteristic and plenty of oxygen functional groups on its surface [36][37][38]. In addition to its superb adsorption property toward the organic pollutants throughout its delocalized π-electron systems [39]. Besides, GO is highly hydrophilic, but also possesses a hydrophobic feature that facilitates the adsorption of the organic contaminants via π-π interaction [40]. Furthermore, it was reported that the incorporation of a low proportion of GO onto the polymer skeleton fosters the mechanical, flexibility, thermal properties and specific surface area of the polymer without affecting its porosity [41]. However, these efficient adsorbents suffer a vast drawback which is their imperfect separation after the adsorption process by the conventional techniques. Notably, shaping GO/ polymer composites in an easily separable form like beads is the most viable solution to solve such an affair [42].
In this perspective, we aimed to fabricate porous GO@ CA beads with a good adsorption behavior and easy separation advantage, exploiting the special durability and mechanical strength of the new CA beads that prevent the GO leaching during the adsorption process. A complete study of the adsorption of MB onto GO@CA beads was executed in a batch mode to assess the adsorption aptitude of the fabricated beads. In addition, the plausible adsorption mechanism of MB onto GO@CA beads was thoroughly explained based on XPS spectra. Besides, a recyclability test was implemented for five sequential cycles. Furthermore, the impact of co-existing cations on the adsorption aptitude of MB was assessed to infer the selectivity of the fabricated beads towards MB dye.

Materials
Cellulose acetate (M.wt. 30,000) was bought from Sigma Aldrich Co. (Germany). Graphite powder, potassium permanganate and sodium nitrate were supplied by Alpha Chemika. Sulphuric acid, hydrogen peroxide and hydrochloric acid were bought from Loba Chemie (India). Ethanol, dimethyl sulfoxide, methanol and methylene blue dye were purchased from Rankem Co. (India). The specifications of MB were summarized in Table S1.

Synthesis of GO
Graphene oxide (GO) was synthesized according to the previously reported modified Hummers method [39]. In brief, accurate amounts of graphite (2 g) and sodium nitrate (1 g) were dissolved at 5 o C in concentrated H 2 SO 4 (100 mL) under continuous stirring. Next, potassium permanganate (10 g) was added to the mixture and followed with a further stirring for 1 h. The temperature of the mixture was raised to 40 ˚C and kept under stirring for 30 min. Thereafter, deionized water (100 mL) was decanted into the mixture and followed by a further rise in the temperature up to 90 o C for 2 h. Finally, deionized water (280 mL) and H 2 O 2 (30 mL) were added to terminate the reaction. The resultant GO was separated and washed using HCl (10%) and distilled water followed by drying for 24 h in an oven at 50 o C.

Synthesis of GO@CA beads
GO@CA beads were fabricated as follows; 1 g CA was dissolved into 5 mL DMSO under mechanical stirring for 30 min. In another beaker, GO (10-25 wt%) was suspended into 5 mL DMSO and sonicated for 30 min. Then, GO was added bit by bit to the CA solution under vigorous stirring. After 30 min, GO-CA composite was added to 250 mL distilled H 2 O drop by drop using a syringe. Finally, the obtained beads were cured for 15 min, collected and washed with distilled H 2 O. Besides, CA beads were fabricated with the same procedure, excluding the step related to adding GO to the polymer matrix. A schematic diagram describes the formulation of GO@CA beads and digital laboratory images for the freshly formulated beads were displayed in Fig. 1.

Characterization
GO, CA and GO@CA beads were characterized by X-ray diffractometer (XRD; MAC Science M03XHF) to inspect their crystal phases. The functional groups of the samples were explored by Fourier transform-infrared spectra (FTIR; Tensor II, Bruker). The thermal stability of the samples was investigated utilizing Thermogravimetric analyzer (TGA; Perkin-Elmer). Moreover, scanning electron microscope (SEM; S4800, Hitachi) was used to inspect the morphology of the samples. The surface charge of GO@CA beads was examined by Zeta-sizer (ZP; Malvern-UK). In addition, the elemental composition of the fabricated beads before and after the MB adsorption was scrutinized by X-ray photoelectron spectroscopy (XPS; Thermo scientific ESCALAB 250Xi VG).

Water uptake study
To investigate the water uptake behaviour of the developed adsorbent beads, an accurate 10 mg of tested sample was soaked in distilled water for 2 h at room temperature. Next, the swollen samples were separated, blotted carefully between two filter papers to eliminate the excess of the adhered surface water droplets and followed by weighing using a closed electronic balance. The percentage of the water uptake (WU; %) was calculated according to the following equation: where, W s and W i represent the final swollen and initial weights of the tested sample, respectively.

Batch adsorption study
A series of experiments were executed to determine the optimum conditions of the MB adsorption onto GO@CA beads. Consequently, to detect the optimum pH, e MB adsorption was implemented at a wide pH range of 3-11. Moreover, to identify the impact of the dose of GO@CA beads on the MB adsorption efficacy, the process was examined using varying doses of the beads in a range of 5-20 mg. Besides, thermodynamics was studied at a temperature range of 25-55 °C, while the adsorption isotherms were studied at the C o range of 50-200 mg/L. Untimely, after each experiment, a sample of the un-adsorbed dye was withdrawn and measured using a spectrophotometer at λ max = 664 nm. The adsorption capacity and the removal % of MB were calculated by the following Equations; (2) Fig. 1 The preparation of GO@ CA beads and digital laboratory images for freshly prepared wet beads.

TGA
TGA profiles (Fig. 2C) elucidate the thermal behaviors of GO, CA and GO@CA beads. All the analyzed samples reveal a mass loss between 30 and 100 °C which is attributed to the water vaporization. The TGA profile of GO signalizes a mass loss between 100 and 208 °C is most likely due to the removal of oxygen-functional groups. While the mass loss beyond 250 °C may be ascribed to the pyrolysis of more function groups [21,48]. Moreover, the TGA profile of CA shows a significant mass loss between 308 and 398 °C which is assigned to the degradation of the polymer chains [49]. The TGA profile of GO@CA beads implies amelioration in the thermal stability compared to the pristine CA.

Zeta potential
ZP measurements implies the role of GO in boosting the surface charges of CA beads since ZP of CA beads and GO@ CA beads are −35.7 and −40.04 mV, respectively, at neutral conditions (Fig. S1). These results may be attributed to the existence of abundant oxygen functional groups in GO (viz., OH and COOH), increasing the negative charges on the GO@CA surface. Thence, the adsorption performance of GO@CA beads towards the cationic contaminants like the noxious MB ought to be more enhanced than the pristine CA beads.

SEM
The SEM image (

Evaluation of the water uptake profile
It is recognized that water uptake property is mainly depending on the existence of hydrophilic functional groups in the matrix. Water uptake behavior of the fabricated composite beads was investigated as shown in Fig. S2. It was observed that increasing the GO content in GO@CA beads matrix from 10 to 20 slightly increased the water uptake value from 15 to 23%. In addition, native CA beads recorded a minimal water uptake value of 11% due to the hydrophobic nature of CA. Notably, the increasing in the water uptake with rising the embedded GO amount in CA beads could be ascribed to where, C o and C t (mg/L) symbolize the initial concentration of MB and the concentration at time t, respectively. m (g) and V (L) symbolize the mass of GO@CA beads and the volume of MB solution, respectively.

Recyclability test
To assess the ability of GO@CA beads to recycle and reuse many times, the beads were used for removing MB for five consecutive cycles as follows; after the adsorption run, GO@CA beads were collected and soaked onto 50 mL ethanol under stirring for 1 h. Then, the used beads were examined in the next adsorption/ desorption cycle.  [44,45]. The FTIR spectrum of GO@CA beads illustrates the distinguishing absorption peaks of GO and CA with a noticeable diminution in the peaks intensity, suggesting the homogeneity between GO and CA.

XRD
Fig. 2B depicts the crystalline phase of GO, CA and GO@ CA beads. The XRD pattern of GO clarifies its discriminative peak at 2θ = 10.25° [46]. While the XRD pattern of CA points out the typical broad peak of CA at around 2θ = 20° [47]. Furthermore, the XRD pattern of GO@CA beads evinces the diffusion of GO onto the CA matrix since their characteristic peaks appeared.

The impact of GO proportions
Fig. 4A elucidates the impact of the proportion of the incorporated GO onto CA beads on boosting the adsorbability of CA beads toward MB. The result clarifies the poor adsorption performance of CA beads since R% and q were 8.58% and 42.27 mg/g, respectively. It was observed a significant enhancement in the R% and q from 28.84% and 77.22 mg/g to 75.63% and 157.96 mg/g with the increase in the GO proportions from 10 to 25%, respectively. This behavior can be the increase in the number of the hydrophilic groups (OH and COOH), which induces the hydrophilicity of beads matrix. Neverthless, further increase in the GO content up to 25% reduced the water uptake value to 21% as a result of the increase in the density of GO@CA beads, which hinders the penetration of water molecules from the outer medium.

The impact of pH
Emphatically, pH is the most effective key parameter on the efficacy of the adsorption process since it dominates the adsorbent surface charges. Hence, the MB adsorption onto GO@CA beads was studied at a wide pH range from 3 to 11. It is apparent from Fig. 4B that the adsorption efficiency of MB dramatically enhanced when pH was raised from 3 to 7 since R% and q increased from 47.65% to 105.96 mg/g to 75.63% and 157.96 mg/g respectively. Then, the adsorption efficacy of MB onto GO@CA is still almost constant when pH exceeded 7. Meanwhile, ZP measurements ( Fig. 4C) elucidate that the surface of GO@CA beads is negatively charged with an enhancement in the amount of the negative charges on the surface by raising the pH medium. Thereby, such an increase in the adsorption aptitude of MB with raising the pH medium may be explained by the strong electrostatic interaction between the negatively charged GO@CA beads and the cationic MB [22]. This result was in line with Hurairah et al. study that also evinced the suitability of the neutral and alkaline media for the adsorptive removal of the detrimental MB [50]. This result may be ascribed to the aggregation of the adsorbent due to the further increase in its dose, resulting in a decrease in the available surface area [51]. On the contrary, this increase in the dose of GO@CA beads enhanced the R% values from 38.07 to 92.29% owing to the existence of abundant binding sites to adsorb the MB molecules [15].

The impact of temperature
Fig. 5B elucidates a diminution in R% and q of MB onto GO@CA beads from 75.60% and 157.96 mg/g to 32.16% and 82.95 mg/g with raising the process temperature from 25 to 55 o C, respectively. This behavior is most likely due to the increase in the thermal energy of the MB molecules with the increase in the system temperature, leading to the desorption of MB from the surface of the beads [52]. This finding suggests the exothermic adsorption nature of the MB molecules onto GO@CA beads. to 308.20 mg/g. This expected behavior may be assigned to explained by the increase in the oxygen-functional groups (viz., OH and COOH) onto the surface of the beads that grasp the MB from the bulk solution via the electrostatic interactions. Furthermore, GO increments the surface area of CA beads, fostering the adsorption performance of the beads [21]. Thence, 25% GO@CA beads were picked out for the rest of the adsorption experiments.

Kinetic study
To investigate the pathway and the rate-determining step of the MB adsorption onto GO@CA beads, the experimental data were examined by PFO and PSO (Fig. 6A). Table S3 summarizes the non-linear expressions of these models. It was deduced from the calculated kinetic parameters (Table 1) that the adsorption of MB onto GO@CA beads obeys PSO since the calculated q values are almost equal to the experimental values. More importantly, the derived R 2 values from PSO are larger than that of PFO. Moreover, the decline in the k 2 with the increase in the MB concentration infers the chemical adsorption process.

Isotherm study
To scrutinize the MB adsorption nature onto GO@CA beads, the equilibrium data were examined by Langmuir, Freundlich, Temkin and D-R isotherm models (Fig. 6B). Table S4 listed the non-linear expressions of the isotherms models.
The calculated isotherm parameters (Table 2) indicate that the studied adsorption process best fits Freundlich model. Also, Temkin model evinces the physical adsorption as well since b T < 80 kJ/mol. While the calculated E value from D-R model suggests the participation of chemical interactions in the adsorption process as E > 8 kJ/mol. Generally, the n value indicates the favorability of the adsorption process; when n < 1, the process is unfavorable, and it is moderately difficult when the n value falls between 1 and 2, while it becomes more favorable when n exceeds 2 [15]. Hence, it was concluded the favorability of the MB adsorption onto the fabricated beads since the n value was found to be 2.04. These results explain the promising value of the maximum MB adsorption capacity onto GO@CA beads reaching 369.85 mg/g. Table 3 represnts a comparison study the increment in the MB concentration that directly boosts the driving forces of the MB molecules to overcome the mass transfer resistance. Consequently, the movement of the MB molecules from their bulk solution to the GO@CA surface increases [17].   Fig. 5 The impact of (A) adsorbent dose, (C) temperature and (C) initial MB concentration on the MB adsorption onto GO@CA beads.
process since the calculated value of ΔHº was negative. In addition, the negative value of ΔSº implied that GO@CA beads are highly ordered at the liquid-solid interface. The negative ΔGº values evinced the feasibility and spontaneity of the MB adsorption process. Conversely, the increase in the process temperature over 318 K rendered the adsorption process feasible and nonspontaneous. This thermodynamic behavior was consistent with Zainal study [63].

Effect of co-existing ions
Certainly, real wastewater contains many types of co-existing ions that compete with the adsorbates for the adsorption sites of the adsorbents, decreasing the efficiency of the adsorption process. Consequently, the impact of interfering cations (viz., Fe 2+ , Zn 2+ , Ni 2+ and Cu 2+ ) on the MB adsorption aptitude was evaluated as elucidated in Fig. 7A. It was monitored a decline in the adsorption efficiency of MB in the existence of these co-existing cations owing to their strong competition with the cationic MB molecules for the binding sites of GO@CA beads. Furthermore, the negative impact of Fe 2+ (q = 149.62 mg/g, R%= 70.80%) > Zn 2+ (q = 150.12 mg/g, R%= 71.29%) > Cu 2+ (q = 151.11 mg/g, R%= 71.24%) > Ni 2+ (q = 152.11 mg/g, R%= 72.24%) on the MB adsorption efficiency (q = 157.96 mg/g, R%= 75.63%). Such results may be attributed to the fact that the higher ionic radii are the lower hydrated species, thereby the migration rate from the bulk solution to the beads surface of Fe 2+ > Zn 2+ > Cu 2+ > Ni 2+ . This result agreed with Tripathy et al. study which also inferred the decrease in the adsorption efficiency of MB in the presence of Fe 2+ > Zn 2+ > Cu 2+ between the adsorption capacity of GO@CA beads and other relevant adsorbents toward MB dye.

Thermodynamic study
Indeed, the process temperature is a dynamic factor that greatly controls the adsorption aptitude [62]. Thereby, it was crucial to investigate the thermodynamic parameters to deduce the nature of the MB adsorption onto GO@ CA beads (Table 4). Change in entropy (ΔSº), change in enthalpy (ΔHº) and change in free energy (ΔGº) were calculated from the summarized equations in Text S4.
It was deduced from van't Hoff plot (Fig. 6C) that the MB adsorption onto GO@CA beads is an exothermic  The N1s-XPS spectrum (Fig. 8D) signalizes the belonging peaks to N-C and N = C of MB at 399.22 and 401.83 eV, respectively. Noteworthy, MB is a large molecule where its length is 13.82 or 14.47 A o and width is 9.5 Ao , so the molecules partially diffuse into the adsorbent pores via the Lewis acid-base interaction. Furthermore, the peaks shift of O1s-XPS spectrum after the adsorption of MB (Fig. 8E) from 532.72 and 532.84 to 532.83 and 532.92 eV, respectively, suggests the electrostatic interaction contribution to the adsorption process. The impact of pH and ZP measurements assert the role of the electrostatic interaction in the MB adsorption process. In addition to the possibility of the H-bond formation between N atoms of MB and H atoms of GO@CA beads and H atoms of MB and O atoms of the beads. The π -π interactions between the aromatic rings of MB and the beads also contributed to the adsorption mechanism as shown in Fig. 9. Furthermore, it was reported in many research papers that the participation of n-π interactions and van der Waals force in the MB adsorption mechanism [65,66]. Such physicochemical interactions could be inferred via the XPS peak shifting of the survey and O1sspectrum after the adsorption process.
> Ni 2+ [64]. More importantly, the impact of co-existing cations was insignificant on the MB adsorption aptitude, reflecting the selectivity GO@CA beads towards MB.

Reusability study
Undoubtedly, the shaping of adsorbents in the beads form has efficiency and economic advantages, thus it was crucial to assess the ability of GO@CA beads to reuse many times after it was proved the efficacy of the beads. The recyclability test (Fig. 7B) points out a decline in R% and q of MB from 75.63% and 157.96 mg/g to 57.61% and 126.86 mg/g after the 5 th cycle, respectively. This result indicates the propitious recyclability of GO@CA, reflecting its potentiality for bountiful applications.

The proposed adsorption mechanism
Kinetic and isotherm studies suggested the participation of both physical and chemical interactions on the MB adsorption process onto GO@CA beads. Therefore, XPS of GO@ CA beads was studied before and after the MB adsorption for understanding the adsorption mechanism. The XPS survey before the MB adsorption (Fig. 8A) illustrates that GO@CA beads consists of O1s and C1s. The C1s spectrum (Fig. 8B) reveals the characteristic peaks of C = O, C-O-C and C-H/C-C at 289.03, 287.53 and 284.74, respectively. Furthermore, the O1s spectrum (Fig. 8C) shows Fig. 9 The proposed adsorption mechanism of MB onto GO@CA beads.