Graphene Oxide/Polyvinyl Alcohol/Fe3O4 Nanocomposite: An Efficient Adsorbent for Co(II) Ion Removal

In this work, an effective nanocomposite-based adsorbent directed to adsorb cobalt (Co2+) ion was successfully synthesized from graphene oxide (GO), polyvinyl alcohol (PVA), and magnetite (Fe3O4) nanoparticles via a coprecipitation technique. The synthesized GO/PVA/Fe3O4 nanocomposite was applied for Co2+ ion removal with the optimized working conditions including 100 min of contact time, 0.01 g of adsorbent dosage, pH of 5.2, and 50°C of temperature. The investigation of adsorption kinetics showed that the adsorption of Co2+ ion onto the GO/PVA/Fe3O4 nanocomposite followed the pseudo-second-order kinetic model with the rate constant k2 being 0.0026 (g mg−1·min−1). The Langmuir model is suitable to describe the adsorption of Co2+ ion onto the GO/PVA/Fe3O4 nanocomposite with the maximum sorption capacity (qmax) reaching 373.37 mg·g−1. The obtained results also indicated that the GO/PVA/Fe3O4 nanocomposite can adsorb/regenerate for at least 5 cycles with a little reduction in removal efficiency. Therefore, we believe that the GO/PVA/Fe3O4 nanocomposite could be used as a potential adsorbent for heavy metal treatment in terms of high adsorption capacity, fast adsorption rate, and recyclability.


Introduction
Nowadays, electronic devices are becoming more and more common in our life, in which rechargeable batteries are an indispensable item in every family and for every individual. However, the reality indicates that the manufacturing of rechargeable batteries, electrodes, gas turbine engines, hard permanent magnets etc. use a huge amount of cobalt (Co) and discharge a lot of cobalt ions (Co 2+ ) into the environment. Inhalation of Co dust may cause adverse respiratory effects, also causing neurological symptoms and cancer in human beings with unknown mechanism [1][2][3].
erefore, there are many kinds of technologies to reduce the concentration of Co 2+ in water pollutants, such as nanofiltration, adsorption, and ion exchange, in which the adsorption process is the best choice because it is cheap and suitable to adapt with a vast range of working conditions to remove Co 2+ ion from aqueous solutions [4][5][6][7][8][9][10]. In order to improve the adsorption efficiency, some advanced nanomaterials have been applied for enhancing the specific surface area of the adsorbent which favors adsorption using carbon nanotubes (CNTs) [11][12][13], activated carbon [14,15], graphite [16], graphene oxide (GO) [17], or reduced graphene oxide [18]. ese nanomaterials have many functional groups on the surface such as −COOH, −OH, and C�O which can be used as an electron-trapping site to attract metal ions or organic materials [11][12][13][14]17]. Recently, GO is widely applied as an adsorbent directed to adsorb heavy metal ions from water because GO has a large surface area (which can be up to 2630 m 2 ·g −1 ) and high water solubility [19][20][21][22]. In addition, GO has abundant oxygen-based groups on its surface such as hydroxyl, carboxylic, carbonyl, and epoxide groups, making GO a material of great interest in adsorption-based technologies of water treatment. GO adsorbents with excellent maximum adsorption capacity (q max ) were reported, such as q max � 198 mg·g −1 for adsorption of Cr(VI) ion [23] and the q max � 46.6 mg·g −1 for adsorption of Cu(II) ion [24]. However, GO is a nano/ micromaterial and it has very low density; therefore, it is difficult to remove GO out of the water after adsorption processes. e hybrids of graphene with magnetic nanomaterials such as Fe 3 O 4 nanoparticles have been used to solve the above problem. Fe 3 O 4 is usually used for water purification due to its safety; in addition, the Fe 3 O 4 is also used to generate magnetic properties for the adsorbent, which makes it easy to be collected after treatment by using an external magnet bar. Yao et al. [25] reported Fe 3 O 4 @ graphene in dye removal with q max of 45.27 mg·g −1 to methylene blue (MB) and 33.66 mg·g −1 to Congo red [25]. Uheida et al. [26] has used Fe 3 O 4 and ɤ-Fe 2 O 3 nanoparticles for the removal of Co 2+ ion. To improve the bonding of Fe 3 O 4 nanoparticles with the GO sheets, a natural polymer or synthetic polymer can be used. In fact, polyvinyl alcohol (PVA) is widely used in the adsorption process because of its nontoxicity, low cost, and chemical stability and having many hydroxyl (-OH) groups [27,28]. Wang et al. [28] have fabricated the GO-PVA composites and showed that GO-PVA can adsorb MB with a q max of 571.4 mg·g −1 . We have reported the use of GO/chitosan/Fe 3 O 4 nanocomposite as a recoverable and recyclable adsorbent for Cr(IV) ion adsorption with easy removal of the GO/chitosan/Fe 3 O 4 composite adsorbent out of the solution by using a magnetic bar and especially high adsorption capacity (q max � 200 mg·g −1 ) for Cr(IV) ion [29]. In this study, we extend the above approach with chitosan being replaced by PVA for synthesis of the GO/PVA/Fe 3 O 4 composite, which was directed to adsorb of Co 2+ ion.  3 O 4 Nanocomposite. 0.1 g of GO was ultrasonicated in 30 ml of distilled water at ambient conditions to have a slurry solution. 0.2 g of PVA was dissolved in 30 ml of distilled water and stirred at 500 rpm at 90°C. When all PVA is dissolved completely and GO is dispersed well, they are mixed with a mixture of Fe 3 O 4 colloid, stirred for 10 minutes and filtered, washed many times with distilled water, and finally dried in an oven at 40°C for 1 day. 2+ Ion Removal. 0.0406 g of CoCl 2 .6H 2 O was dissolved in 100 mL distilled water to obtain the stock solution of Co 2+ ion (100 mg·L −1 ). e stock solution was diluted to the desired solution. 0.01 g of GO/PVA/Fe 3 O 4 nanocomposite as an adsorbent was added into 20 mL of the solution containing Co 2+ ion; then, the mixtures were agitated at 30°C and pH 7 for 100 min. e residue Co 2+ ion concentration in the solution was analysed by the spectrophotometric method (described in Section 3.5). e adsorption capacity, q (mg·g −1 ), and the percentage removal (H, %) were calculated by the following equations:

Batch Adsorption Experiments for Co
where C 0 and C e are the initial and equilibrium concentrations of Co 2+ solution (mg·L −1 ), respectively; C is the Co 2+ ion concentration at time t; V is the volume of the sample solution used for the experiment (mL); and m is the weight of the adsorbent (g). e adsorption of Co 2+ ion onto GO/ PVA/Fe 3 O 4 was studied as a function of contact time, mass of adsorbent influent, pH of the solution, and temperature. e used GO/PVA@Fe 3 O 4 adsorbent was recovered by immersing it into 0.1 N NaOH solution for 1 day, and then, it was rinsed with distilled water and dried at 40°C in an oven for 12 h to obtain the regenerated GO/PVA@Fe 3 O 4 adsorbent, which was reused to adsorb Co 2+ ion.

Determination of Co 2+ Ion Concentration.
e Co 2+ ion concentration after adsorption process is determined by the spectrophotometric method with the aid of many complexing reagents, which had been developed recently to monitor Co 2+ concentration, with fast response, high sensitivity, and easy preparation compared to other methods. e common complexing reagents used include ninhydrin (optimum pH is 8.2; the complex is stable in 30 min) [30], 2benzoylpyridine-4-phenyl-3-thiosemicarbazone [31], 2pyridine carboxaldehyde isonicotinyl hydrazone (pH 9; the apparent molar adsorptivity is 7.1.10 4 L·mol −1 ·cm −1 ) [32], dehydroacetic acid oxime (pH 5.8) [33], and 5-[3-(1,2,4triazolyl-azo]-2,4-dihydroxybenzaldehyde [34]. In our work, we used NH 4 SCN as a ligand in acetone solution and we have discovered that it can be stable up to more than 7 days. A typical procedure is followed: 1 ml of 18% HCl solution is added into 10 ml of residue solution at RT. en, 0.5 ml of NH 4 SCN saturated solution is added into the above mixture and the solution is mixed well. After that, 20 ml of concentrated acetone is added into this mixture and the solution will change from pink color to blue color in the following reaction: 2 Journal of Analytical Methods in Chemistry e absorbance of the mixture is then measured with an Agilent 8453 UV-Vis spectrophotometer, and the calibration curve is obtained ( Figure S1). Each sample was measured in duplicate.
2.6. Characterizations. XRD patterns of GO and GO/PVA/ Fe 3 O 4 were obtained on D8 Advance, Brucker ASX, operated at a CuK α wavelength of 1.542 _ A in the range of 2θ � 5 to 70°at the room temperature. UV-Vis spectra were measured with an Agilent 8453 UV-Vis spectrophotometer system. e fracture surfaces of GO and GO/PVA/Fe 3 O 4 were observed using a Hitachi S4500 Scanning Electron Microscope (SEM). e infrared (IR) spectra were recorded on a Nicolet FT-IR Spectrometer model 205 with KBr pellets in the region from 500 cm −1 to 4000 cm −1 . Specific surface area and pore size distribution of the prepared GO/PVA/Fe 3 O 4 sample were evaluated using low-temperature nitrogen adsorption isotherm by Brunauer-Emmett-Teller (BET) and Barrett-Joyner-Halenda (BJH) methods on the Tristar II plus System (Micromeritics, USA).

Characterization of the GO/PVA/Fe 3 O 4 Nanocomposite.
e XRD patterns of GO and GO/PVA/Fe 3 O 4 (Figure 1(a)) show a diffraction peak at 2θ � 10°, which is assigned to the crystalline of GO (curve A) with (001) reflection indicating that the oxygen functionality existence increases the distance between graphene layers. Using Bragg's Law and Scherrer equation for this sharp peak, it is revealed that the interlayer space is about 0.885 nm and the number of layers in GO is 5. In case of GO/PVA/Fe 3 O 4 (curve b), this peak disappears due to a very low content of GO in the sample; however, the characteristic peaks of Fe 3 O 4 clearly appeared at 2θ � 30°, 35°, 57°, and 63°corresponding to the reflection of (220),  (Figure 1(g)) exhibits type IVa hysteresis loops by IUPAC, which is specific to mesoporous materials with a pore width range from 4 to 50 nm [35]. e BJH pore size distribution of GO/PVA/Fe 3 O 4 sample (Figure 1(h)) shows the main pore diameters to be less than 7 nm, which is in agreement with the shape of the hysteresis loop above (Figure 1(g)). e BET specific surface area and BJH average pore width of the synthesized GO/PVA/Fe 3 O 4 sample is summarized in Table 1. Figure 2(a) shows that Co 2+ ion adsorption capacity increased rapidly when contact time was from 3 to 100 minutes, with about 60% of the Co 2+ removed, and thereafter, the adsorption capacity has a constant trend. e rapid uptake within 100 min was due to the large surface area, the presence of various oxygen functional groups of GO, and PVA that creates an electrostatic interaction with Co 2+ ion. After that, the adsorption sites of the adsorbent were filled with Co 2+ ions so the rate of adsorption becomes constant. e contact time here longer than that in other reported materials [36] can be attributed to higher of Co 2+ initial concentration and lower used adsorbent dose (100 mg·L −1 , m � 0.01 g) as well. e influence of adsorbent dosage was evaluated by changing the mass of GO/PVA/Fe 3 O 4 adsorbent from 0.0123 g to 0.034 g for treatment of 20 mL of Co 2+ solution, and the obtained results show that with the increasing mass of the adsorbent, the adsorption capacity decreases and the optimal amount of absorbent is 0.0123 g with an adsorption capacity q e is 17.63 mg g −1 (Figure 2(b)). Figure 2(c) reveals that the q e increased with the increase in temperature, which suggested that the adsorption of Co 2+ ion onto the GO/PVA/Fe 3 O 4 adsorbent may be favored by high temperature and therefore the optimal temperature for this process was selected at 50 C. e influence of pH on the adsorption process was evaluated with pH change from pH 2 to pH 7 because at higher pH (pH > 7), Co 2+ ion can be agglomerated as a Co(OH) 2 precipitate [37,38]. As shown in Figure 2(d), Co 2+ ion removal was 61.7% with a q e of 121 mg·g −1 at pH 2 (curve A) and the removal was about 64.0% with a q e of 127.3 mg·g −1 at pH 5.2 (curve B), and the UV-Vis spectra at equilibrium time are shown in Figure S2. e obtained results can explain that the high concentration of H + ion (at low pH value) led to the competition between positive charge ions to attach with negative charge oxygen functional groups on GO and PVA. Meanwhile, at higher pH, the concentration of H + ions decreases so there is less competition, and the result is the adsorption increase. erefore, the optimal pH for adsorption Co 2+ ion onto GO/PVA/ Fe 3 O 4 was pH 5.2. Effect of K + ion as an interference to the adsorption of Co 2+ onto GO/PVA/Fe 3 O 4 was also tested (  interfere with the adsorption efficiency of Co 2+ ion, even when the concentration of interfering K + ion was 10-80 times higher than the Co 2+ ion concentration, which is completely consistent with the previous report [39].

e Kinetics of Co 2+ Ion Adsorption onto GO/PVA/Fe 3 O 4
Nanoadsorbent. In this work, two kinetic models including the pseudo-first-order and the pseudo-second-order    kinetic models were analysed, which can be represented as follows: ln q e − q t � ln q e − k 1 t, where q e is the adsorption capacity at equilibrium, q t is the adsorption capacity at time t (min), and k 1 is the pseudofirst-order rate constant of adsorption (min −1 ), and k 2 is the pseudo-second-order rate constant of adsorption (g·mg −1 ·min −1 ). As proved in the previous study [40], equation (2) can be written as follows: where C 0 and C t are the initial concentration and concentration of Co 2+ ion in solution, respectively, A o and A e are the initial and equilibrium absorbance; k 1 was calculated from the slope of the plot of ln (A t − A e ) vs. t (Figure 3(a)); and k 2 was calculated from the slope of the plot of t/q t vs. t. (Figure 3(b)). Based on comparison of the correlation coefficient (R 2 ) of two plots, it can be seen that the adsorption of Co 2+ ion onto the GO/PVA/Fe 3 O 4 adsorbent was fitted to the pseudo-second-order kinetic model than the pseudofirst-order kinetic model. e rate constant k 2 was fitted at 0.0026 g·mg −1 ·min −1 .

Investigation of the ermodynamic Parameters of the Adsorption Process.
e thermodynamic parameters of the adsorption process such as enthalpy change (ΔH 0 ), entropy change (ΔS 0 ), and Gibbs free enthalpy change (ΔG 0 ) of reaction are calculated following the second law of thermodynamics, which have been described by the following equations: where R is the gas constant (R � 8.314 J·mol −1 ·K −1 ), K c is the equilibrium constant of chemical reaction, and T is the absolute temperature (K). ΔH 0 and ΔS 0 can be calculated from the slope and the intercept on the plot of lnK C vs. 1/T following equation (4) and the obtained results are given in Table 3. e negative value of ΔG 0 indicates that the process is spontaneous for all evaluated temperatures. e positive value of ΔH 0 confirms that the adsorption is endothermic and favors at high temperature. e positive value of ΔS 0 shows the increase in the randomness of the adsorption on GO/PVA/Fe 3 O 4 for Co 2+ ion.

Adsorption Isotherm.
e Langmuir model (equation (5)) assumes a monolayer adsorption onto the homogeneous surface, and there is no transmigration of adsorbate on the surface plane. Meanwhile, the Freundlich model assumes a multilayer adsorption onto the heterogeneous surface (equation (5)). where q max (mg·g −1 ) is the maximum adsorption capacity of Co 2+ ion onto the GO/PVA/Fe 3 O 4 adsorbent; K L, K F are the Langmuir constant and Freundlich constant, respectively; and n is a constant. Experimental results following the Langmuir and Freundlich adsorption isotherm are shown in Figure 4, and the fitting of the Langmuir and Freundlich constants is given in Table 4. It can be seen that, with the higher correlation coefficient, the Langmuir model fitted well for the adsorption of Co 2+ ion onto GO/PVA/Fe 3 O 4 ( Table 4). e maximum monolayer adsorption capacity q max is 370.37 mg·g −1 and K L is 0.0122. Compared to the other adsorbents in the literature for Co 2+ ion removal (Table 5), the obtained result in our work is so high and impressive, which can be attributed to a very large surface area of GO, facilitating the adsorption process.

Recordation and Regeneration
Studies. SEM images of GO/PVA/Fe 3 O 4 adsorbent before ( Figure 5(a)) and after Co 2+ adsorption process ( Figure 5(b)) are not significantly different excepting several clusters of particles appearing on the surface of adsorbent after the adsorption process, and the surface of the GO/PVA/Fe 3 O 4 adsorbent after adsorption process is less porous than that before. ese observations can be attributed to the presence of adsorbed Co 2+ ion onto the GO/PVA/Fe 3 O 4 surface, which was confirmed by the EDS analyses. It can be seen that there was no cobalt element (0 wt.%) on the EDS spectrum of GO/PVA/Fe 3 O 4 before adsorption ( Figure 5(c)); meanwhile, with the sample after Co 2+ ion adsorption, the cobalt element reached 4 wt.% ( Figure 5(d)), in which Co 2+ ions were adsorbed onto the GO/PVA/Fe 3 O 4 adsorbent.
As shown in FIgure S3, the regenerated GO/PVA/Fe 3 O 4 adsorbent can adsorb Co 2+ ion with only a small decrease in adsorption capacity between the 1st cycle and the 5th cycle. e remaining adsorption efficiency of the 5th cycle was about 86% of the 1st cycle ( Figure S3), which implies that the GO/PVA/Fe 3 O 4 material has a good stability and a high

Conclusion
As described in this study, GO/PVA/Fe 3 O 4 nanocomposite as an effective adsorbent has been simply synthesized via a coprecipitation technique and it was directed to capture Co 2+ ions from an aqueous solution via an adsorption process. e adsorption process of Co 2+ ion onto the GO/PVA/Fe 3 O 4 adsorbent was evaluated by studying the effects of adsorbent dose, the solution pH, and temperature. At optimized adsorption conditions, this process followed the pseudo-second-order kinetic model and the adsorption isotherm was fitted to the Langmuir model with the maximum adsorption capacity to Co 2+ ion estimated at 370.37 mg·g −1 . e GO/PVA/Fe 3 O 4 adsorbent can be regenerated for at least 5 cycles with the remaining adsorption efficiency after the 5th cycle being 86% compared to the first cycle. ese findings indicate that the GO/ PVA/Fe 3 O 4 nanocomposite can be considered as a good candidate for the removal of Co 2+ ion from aqueous solutions.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that they have no conflicts of interest.
Acknowledgments is work was supported by the Ministry of Education Project, under code number B2020-BKA-15. Figure S1(a) UV-Vis spectra of various concentrations of Co 2+ solution and (b) the calibration curve for Co 2+ concentration measurement. Figure S2