Exploration of the Cs Trapping Phenomenon by Combining Graphene Oxide with α-K6P2W18O62 as Nanocomposite

A graphene oxide-based α-K6P2W18O62 (Dawson-type polyoxometalate) nanocomposite was formed by using two types of graphene oxide (GO) samples with different C/O compositions. Herein, based on the interaction of GO, polyoxometalates (POMs), and their nanocomposites with the Cs cation, quantitative data have been provided to explicate the morphology and Cs adsorption character. The morphology of the GO-POM nanocomposites was characterized by using TEM and SEM imaging. These results show that the POM particle successfully interacted above the surface of GO. The imaging also captured many small black spots on the surface of the nanocomposite after Cs adsorption. Furthermore, ICP-AES, the PXRD pattern, IR spectra, and Raman spectra all emphasized that the Cs adsorption occurred. The adsorption occurred by an aggregation process. Furthermore, the difference in the C/O ratio in each GO sample indicated that the ratio has significantly influenced the character of the GO-POM nanocomposite for the Cs adsorption. It was shown that the oxidized zone (sp2/sp3 hybrid carbon) of each nanocomposite sample was enlarged by forming the nanocomposite compared to the corresponding original GO sample. The Cs adsorption performance was also influenced after forming a composite. The present study also exhibited the fact that the sharp and intense diffractions in the PXRD were significantly reduced after the Cs adsorption. The result highlights that the interlayer distance was changed after Cs adsorption in all nanocomposite samples. This has a good correlation with the Raman spectra in which the second-order peaks changed after Cs adsorption.


Introduction
The destruction of the reactor cores of the Fukushima Daiichi Nuclear Power Plant (FDNPP) ten years ago has had a large influence on the environment. Many researchers predicted the quantity of released material [1] and modeling of the radiation fallout immediately [2]. It is important to consider the radioactive isotopes delivered to the agriculture area [3], tap water [4], and forest area [5]. The subject of considerable concern for the people is the possibility of the radiation exposure to their body. Both the external exposure and internal exposure have radiological consequences that can cause cancer [6,7]. Therefore, it is an essential task to treat and decontaminate the contaminated environment.

Materials
All raw chemicals were reagent grade and were used without purification, and distilled water was used in all experiments. Graphite, Na 2 WO 4 ·2H 2 O, CsCl, and methanol were purchased from Sigma Aldrich Chemistry. NaNO 3 , H 2 SO 4 , HCl, H 3 PO 4 , and KCl were purchased from Fuji Film Wako Pure Chemical Corporation. KMnO 4 and H 2 O 2 were purchased from the Tokyo Chemical Industry.

Synthesis of Graphene Oxide
All GO samples were synthesized using the modified Hummers method [27]. A pre-cooling procedure was implemented by keeping all the reagents and solvent in the refrigerator for one night before the synthesis process [28]. In order to start the synthesis process, first, a graphite (1.0 g) and NaNO 3 (1.5 g) were put into the vial. Then, a concentrated H 2 SO 4 (50 mL) was added carefully. We stirred the mixture for 30 min at room temperature. In this step, the color was blackish green. Then, we put the mixture in the refrigerator for 30 min to keep it in the cooling condition below 10 • C. After that, going to the critical step, we added KMnO 4 4.0 g slowly and carefully. Then, we stirred it for about 20 min and then set the temperature at 35 • C under stirring for 2 h. After this reaction, the color of the solution became brown. We diluted the mixture with 80 mL of distilled water with stirring for 15 min. Then, we continued stirring for the other 30 min at room temperature. We added 150 mL of distilled water slowly and carefully with stirring for 15 min. After that, 12 mL of 30% H 2 O 2 was dropped carefully to the solution. A large number of bubbles were released. The solution became golden yellow. The solid particle was filtered, washed using 5% HCl, and followed by using distilled water. We dried the solid at 60 • C for 24 h. In order to distinguish each GO sample, we termed these GO samples by index number, which have different amounts of C based on elemental analysis. There are two types of GO sample based on C/O composition. The GO c70 and GO c72 have 70-72 wt % C element, while GO c39 and GO c40 have ≈ 39 wt % C element.

Synthesis of α-K 6 P 2 W 18 O 62
The α-K 6 P 2 W 18 O 62 was synthesized according to the modified Nadjo method [29]. The synthesis route started with dissolving the starting material, Na 2 WO 4 ·2H 2 O 15 g into 17.5 mL of distilled water. We vigorously stirred for 15 min until the cloudy solution became limpid. This was followed by an acidification process: HCl 4M (12.5 mL) was added dropwise for 45 min carefully. In this process, white precipitate was formed. In order to avoid agglomeration and to disperse homogenously during the acidification process, the vigorous stirring was performed. We checked the pH to be around 6-7. Then, 2M H 3 PO 4 (12.5 mL) was added, and we checked the pH to become around 2-3. The solution was refluxed for 24 h; then this solution was cooled at room temperature for a while. Then, 7.5 g of KCl was added to the solution. Afterwards, this solution was filtered off by paper filtration, and the crude material was obtained. Then, we dissolved it into 32.5 mL of distilled water, and the solution became limpid. The next reflux process was performed at 80 • C for 72 h. Then, it was cooled at room temperature. The last step was keeping the solution in the refrigerator, and after 24 h, the well-behaved crystals of α-K 6 P 2 W 18 O 62 were produced (6.5 g).

Synthesis of GO-POM Composite
The composite sample was distinguished into eight types of composite. Each composite was produced by incorporating the GO sample (GO c70 , GO c72 , GO c39 , and GO c40 ) with Dawson-type POM. The GO-POM composite was generated in two different concentration ratios (concentration ratios of GO:POM: 1:8 and 4:1). The constitution is summarized in Section 3.7.2. These treatments (different concentration ratios of GO:POM) were employed in order to investigate the role of GO and Dawson-type POM in the Cs adsorption and to know the proper composition. In this synthesis process, mixed solvent was used with the ratio 1:1 (distilled water/methanol). The GO aqueous dispersion in the first concentration condition (1:8) was prepared by dispersing 1 g of GO into the solvent (100 mL) with sonification treatment for 5 min. Then, we prepared the POM solution by dissolving 1 g of α-K 6 P 2 W 18 O 62 into the solvent 12.5 mL. Afterwards, we poured the POM solution into the GO suspension vial, added diluted HCl to adjust the pH in the range 2-3, and then vigorous stirred for 24 h. In the second concentration condition (4:1), we prepared the GO solution by dispersing GO (0.5 g) into the solvent (25 mL) with a sonification treatment for 5 min. The POM solution was prepared by dissolving the α-K 6 P 2 W 18 O 62 0.5 g into the solvent (100 mL). The POM solution was poured into the GO suspension; then, we added diluted HCl to adjust the pH in the range 2-3 and after that stirred for 24 h. Finally, the solid material was collected by using filter paper and dried at 90 • C for 24 h. Then, the composite product was indexed based on the value of the C element of GO (c70, c72, c39, and c40 due to elemental analysis) and the concentration ratio of GO/POM (1:8 and 4:1) in order to distinguish each composite's name, as follows: [GO 70 41 . Contant's group reported the hydrolytic stability of Dawson-type POM depending on the acidity of the solution. That is, Dawson-type POM was stable at pH lower than 6, and when the pH increased above 6, the formation of lacunary species occurs [30]. In the case of our system, preparation and adsorption experiments were performed at pH lower than 5. Therefore, the Dawson-type POM was stable in the present solution.

Adsorption Experiment
The cesium adsorption was conducted as below. First, 0.08 g of each sample (POM, GO samples, GO-POM composites) was dispersed into 30 mL of distilled water with sonification treatment for 5 min. Second, 0.6 g of CsCl (3.6 mmol) was added to the solution. Then, the mixture was stirred for 24 h.
The performance of Cs adsorption was analyzed by using ICP optical emission spectroscopy. The solid adsorbent was separated from the solution by using paper filtration. The measurement was conducted in duplicate in each sample solution. The adsorption efficiency was calculated by Equation (1), and the adsorption capacity was calculated by Equation (2).
Adsorption e f f iciency (% Ads.e f f . where C 0 (mmol/L) is the initial concentration of Cs, and C t (mmol/L) is the remained concentration of Cs in the solution after treatment. W (g) is the dry weight of the adsorbent used in the adsorption experiment. Table 1 shows the results of elemental analysis of GO c70 , GO c72 , GO c39 , and GO c40 . It can be seen that the GO c70 and GO c72 have 70-72 wt % C element, while GO c39 and GO c40 have ≈39 wt % C element. In the synthesis process, there is one thing that might make the sample a bit different. In the treatment of KMnO 4 , especially for GO c39 and Materials 2021, 14, 5577 5 of 20 GO c40 , the KMnO 4 was treated more carefully (one minute stop, one minute go). It was shown that the degree of oxidation was controlled by the controlled oxidation process. In particular, it can be assumed that there is a difference in structure between the two types of GO. GO c70 and GO c72 indicating few oxygen defects in the structure have an important contrast with GO c39 and GO c40. However, deeper characterization is still required, since the oxygen defect is not only the case. As emphasized in the previous literature, there is no evident relation between the oxidation degree and ordered domain that is exhibited in the Raman spectrum [31].

FT-IR Spectra
Figures S6-S8 (see supporting information) show the FT-IR spectra of GO and Dawsontype compounds. The Dawson-type POM was successfully produced by the Nadjo synthesis route. The IR absorptions of the present α-K 6 P 2 W 18 O 62 are similar with those of the literature [29]. All the GO samples exhibit a broad absorption band of O-H stretching vibration around 3400 cm −1 , C=C bond stretching at around 1620 cm −1 , C=O vibration at around 1720 cm −1 , C-O vibration band at around 1420 cm −1 , and epoxides at 1200-1250 cm −1 . In addition, the vibration band at around 830 cm −1 -850 cm −1 was detected especially in GO c39 and GO c40 samples. This band might be assigned as the C-Cl vibration band as reported in the other work [32]. However, the C-Cl vibration band is unexplained, because in this work, there is no other additional treatment that refers to the addition of Cl. Since the GO is a nonstoichiometric compound, the structure analysis is quite difficult, and the variety of the composition depends on the synthesis condition. It can be assumed that the remaining Cl impurity is coming from laboratory glass material. The difference in the degree of oxidation was revealed between GO c70 , GO c72 , and GO c39 , and GO c40 ; i.e., an epoxide signal was seen in the GO c39 and GO c40 samples.

Morphology of GO
TEM imaging and SAED diffraction are used to draw the structural model in each GO sample, as shown in Figure 1. For TEM image measurement, 200 kV was applied in each sample. Under this condition, the morphologies of the GO samples were shown to be lamellar, with wrinkles and overlaps, and they were also found to be stable. The image indicates that the some GOs had been exfoliated. This is in agreement with the previous research in which both wrinkles and overlaps are essential features of GO-based bulk material [33]. The sharp spots in an SAED pointed out the crystalline region in GO. Especially GO c72 shows a typical diffraction pattern representing the crystalline nature of the GO sample. The present SAEDs are in agreement with the previous work that presented the amorphous region (see Figure 1(A1,C1)) and crystalline region (see Figure 1(B1)) in the GO sample [34]. Particularly, the amorphous zone is associated with the presence of the sp 2 /sp 3 carbon cluster (see Section 3.7.2) in all the GO samples [35]. In the other imaging using SEM measurement, it is shown in Figure 1d,e that the GO sample consists of several layers (see Table S2, supporting information). Previous work emphasized that the morphology and quality of the multilayers of GO depend on several factors such as the source of graphite, degree of oxidation, and reaction condition [36]. All these GO samples are aligned with a few layers of GO in the previous work [37].
using SEM measurement, it is shown in Figure 1d,e that the GO sample consists of several layers (see Table S2, supporting information). Previous work emphasized that the morphology and quality of the multilayers of GO depend on several factors such as the source of graphite, degree of oxidation, and reaction condition [36]. All these GO samples are aligned with a few layers of GO in the previous work [37].

Morphology of GO-POM Composite and Its EDS Map
The SEM image of the GO-POM nanocomposite and its EDS map are shown in Figures 2, S1 and S2 (see supporting information). Figure 2a shows the SEM of [GO70POM]41, and Figure 2b-d show the W, O, C element distribution, respectively. Tungsten is dominantly originated from the α-K6P2W18O62 particle, and carbon is coming from the GO. The result suggests that the tungsten, oxygen, and carbon elements were homogeneously distributed in the GO-POM nanocomposite, showing the successful combination between GO and POM [38]. This phenomenon was also confirmed by TEM measurement.

Morphology of GO-POM Composite and Its EDS Map
The SEM image of the GO-POM nanocomposite and its EDS map are shown in Figure 2 and Figures S1 and S2 (see supporting information). Figure 2a shows the SEM of [GO 70 POM] 41 , and Figure 2b-d show the W, O, C element distribution, respectively. Tungsten is dominantly originated from the α-K 6 P 2 W 18 O 62 particle, and carbon is coming from the GO. The result suggests that the tungsten, oxygen, and carbon elements were homogeneously distributed in the GO-POM nanocomposite, showing the successful combination between GO and POM [38]. This phenomenon was also confirmed by TEM measurement.

Powder X-ray Diffraction Pattern (PXRD)
Powder X-ray diffraction (PXRD) measurement was carried out to characterize the GO, Dawson-type POM, and GO-POM nanocomposite structures. As shown in Section 3.7.1, Figures S4 and S5 (see supporting information), the PXRD patterns of the GO c70 , GO c72 , GO c39 , and GO c40 show diffractions at 2θ = 11.58 • , 8.66 • , 8.91 • , and 9.13 • , respectively. The d-spacing increases from 0.34 nm of graphite spacing (002) to 0.77-1.02 nm, which corresponds to the typical diffraction peak of the GO nanosheet (Table S2, see supporting information). The interlayer distance was calculated by Bragg's law, as shown below: where d is the distance between the layers of GO, θ is the diffraction angle, and λ is the wavelength of the X-ray beam (λ = 0.154 nm).
Although the d-spacing of GO c70 is shorter than those of GO c72 , GO c39 , and GO c40 , it is suggested that oxygen-containing functional groups (epoxide, hydroxyl on the basal plane, and carboxyl groups on the edge of the basal plane) successfully formed in different degrees of oxidation. This is the reasonable explanation to the enhancement of the interlayer spacing of GO [39].

Powder X-ray Diffraction Pattern (PXRD)
Powder X-ray diffraction (PXRD) measurement was carried out to characterize the GO, Dawson-type POM, and GO-POM nanocomposite structures. As shown in Section 3.7.1, Figures S4 and S5 (see supporting information), the PXRD patterns of the GOc70, GOc72, GOc39, and GOc40 show diffractions at 2θ = 11.58°, 8.66°, 8.91°, and 9.13°, respectively. The d-spacing increases from 0.34 nm of graphite spacing (002) to 0.77-1.02 nm, which corresponds to the typical diffraction peak of the GO nanosheet (Table S2, see supporting information). The interlayer distance was calculated by Bragg's law, as shown below: d = λ/2 sin (θ) = 0.154 nm/2 sin (θ) (3) where d is the distance between the layers of GO, θ is the diffraction angle, and λ is the wavelength of the X-ray beam (λ = 0.154 nm).
Although the d-spacing of GOc70 is shorter than those of GOc72, GOc39, and GOc40, it is suggested that oxygen-containing functional groups (epoxide, hydroxyl on the basal plane, and carboxyl groups on the edge of the basal plane) successfully formed in different degrees of oxidation. This is the reasonable explanation to the enhancement of the interlayer spacing of GO [39].
The diffraction peaks of all samples were slightly shifted by forming the composite (Figures S4 and S5 (see supporting information)). That is, the d-spacing decreased by forming the composite, except for [GO70POM]41 (Table S2, see supporting information). The result of [GO70POM]41 might agree with the previous work that might be assumed as a consequence of a partial intercalation of the POM species and random arrangement of the carbon layers [40]. The results except for the [GO70POM]41 result revealed that the α- The diffraction peaks of all samples were slightly shifted by forming the composite (Figures S4 and S5 (see supporting information)). That is, the d-spacing decreased by forming the composite, except for [GO 70 POM] 41 (Table S2, see supporting information). The result of [GO 70 POM] 41 might agree with the previous work that might be assumed as a consequence of a partial intercalation of the POM species and random arrangement of the carbon layers [40]. The results except for the [GO 70 POM] 41 result revealed that the α-K 6 P 2 W 18 O 62 species are not embedded in between the GO sheets; however, it could be assumed that the POM species decorated on the surface of GO, as is assumed in Figure  S3 (see supporting information). It is already known that the molecular size of the α-K 6 P 2 W 18 O 62 species is larger (ca. 1.03 nm × 1.5 nm) [20] than the d-spacing of each GO sample (0.77-1.02 nm).
In addition, we could see in Table S2 (see supporting information) the out of plane crystallite size of GO samples. The size was calculated by using the Debye-Scherrer equation: where D is the average out of plane crystallite size (nm), λ is the wavelength of the X-ray beam, β is the FWHM (full width at half maximum), θ is the XRD peak position. The in-plane crystallite size of the GO samples and the average number of graphene oxide layers (n) per domain were also estimated by using PXRD data. The following general formula is described in [41]: where D is the average out of plane crystallite size (nm), d is the interlayer distance, λ is the wavelength of the X-ray beam, β is the FWHM (full width at half maximum), θ is the XRD peak position, and n is the number of layers of the GO sample. These calculations provide crucial information regarding the average crystallite size (reflecting the enhancement or reduction of the graphitic zone and the grain boundaries or lateral defect formation) and the average number of GO layers (see Table S2, supporting information).

Raman Spectroscopy
Raman spectroscopy was used to characterize the GO material, and also, it provides the structural information in detail to draw its quality [42]. From the spectra, we could see the D bands located at 1356 cm −1 , 1352 cm −1 , 1352 cm −1 , and 1347 cm −1 for GO c70 , GO c72 , GO c39 , and GO c40 , respectively (see Figure S9 supporting information). This indicates the extensive oxidation of graphite and extended amounts of the sp 3 -hybridized carbon atoms [43]. The G bands located at 1586, 1581, 1590, and 1590 cm −1 for GO c70 , GO c72 , GO c39 , and GO c40 , respectively, correspond to the sp 2 -hybridized carbon atoms in the hexagonal framework. The G band is inherently with the 2D band (the second-order Raman modes) that appears at 2700 cm −1 . The critical thing to fully and accurately characterize the GO sample is the consideration of the second-order Raman modes [44]. The area ratio of the 2D band also has a linear correlation with the hole mobility [45]. In addition, the previous work has explained the packing mechanism of multiple layers GO (bi-, tri-layers) by STEM-ADF image simulation. It showed that the disorder of the graphene sheet and the roughness have been created by the random covalent attachment of oxygen on the top and bottom surfaces. It influences the lattice distortion and breaks the symmetry of the system [46]. Therefore, it can be predicted that the different C/O ratio of GO samples in our work is sufficient to influence the material structure of the GO-POM composite and the ability to adsorb the Cs.
As presented in Table S1 (see supporting information), the intensity ratio of I D /I G of each original material (GO c70 , GO c72 , GO c39 , and GO c40 ) are 0.93, 0.91, 0.89, and 0.88, respectively. These values are correlated with the different compositions of the C/O ratio on each GO sample that are already discussed in the previous section. The intensity ratio (I D /I G ) tends to increase by forming a composite. Therefore, we may conclude that the anchored process of POM on the GO has affected the structure of GO. In this case, also, there was an opposite trend for GO c70.

Adsorption Efficiency and Adsorption Capacity of GO and POM
The Cs adsorption capacities of GO c70 , GO c72 , GO c39 , and GO c40 and POM are summarized in Table 2. Those of GO-POM nanocomposites are presented in Table 3 (see Section 3.7.2). The GO samples were classified into two types according to the elemental analysis. The GO c70 and GO c72 have 70-72 wt % C element, while GO c39 and GO c40 have ≈39 wt % C element. As shown in Table 2, the result of the present work pointed out that there is not much difference in the Cs + adsorption by POM, GO c70 , GO c72 , GO c39 , and GO c40 among the samples. The role of different C/O ratios in capturing Cs cannot be revealed. It is also already shown by other researchers that the C/O ratio of GO does not have systematic correlation with the ion exchange capacity [43].  Furthermore, in this work, the interaction between GO c70 and the Cs cation occurred immediately less than one minute after the stirring process (see Figures S14 and S15, supporting information). A brown milky coagulation was formed. The same behavior was observed between GO c72 and Cs; however, there was a difference for GO c39 : the brown milky coagulation was formed in a small quantity, and the milky coagulation was not formed for GO c40 . The difference might be related to the variance in C/O ratio between GO c70 , GO c72 , and GO c39 , GO c40 . Moreover, it was explained that by using a surface functional group, the GO can attract the Cs cation by forming inner-sphere complexes [47]. The metal complexing in the oxidized GO strongly encourages the aggregation behavior and can be quantified by the critical coagulation concentrations (CCC) [48]. TEM measurement shows that there are many Cs clusters in the surface of GO c70 (see Figure 3c). Especially for this phenomenon, we also considered that the Cs cluster may include Al impurity (see Figure 3d). It might come from laboratory glass material, as reported by other work [31]. The black spots also can be seen in the [GO 40 POM] 41 , which consists of GO c40 /POM. However, the black spots are not uniformly found all over the surface area. It is theoretically explained that most of the surface reactions eventuate only on the active site and do not happen uniformly over the surface area [49]. As proposed by other work, the edge site of graphene (carbene-and carbyne-type carbon atoms) has a significant role for adsorption on the surface of sp 2 -hybridized carbon [50]. It is also confirmed by previous study that some ions or molecules dissolved in water have a good penetration rate through the graphene nanocapillaries until the physical size of the ions or molecules exceeds the critical one [51]. A highly oxidized domain and the hydrophobic domain in GO have different roles. The highly oxidized domain would be responsible for rapid water transport in the interlayer of GO, while the hydrophobic domain would serve a network of capillaries [43]. Another researcher also considered the polarizability of graphene in order to estimate the orientation of interfacial water [52], and the interfacial water can be used to observe the surface adsorption behavior [53]. Based on these analyses, it can be assumed that the interfacial water may support the Cs interaction with GO. The other important evidence of Cs adsorption is also visible in PXRD measurement. The sharp and intense diffraction peaks were reduced by introducing Cs (Figures 4 and  S5 (see supporting information)). The broad diffraction of GOc70, GOc72, GOc39, and GOc40 showed the non-uniform surface complexation with the Cs cation. PXRD and TEM show the formation of the Cs cluster. This might correspond to the origin of GO acidity, which contributes to the high cation exchange capacity (CEC). The GO acidity is made by oxygen functional groups, and the main moieties responsible for the acidic properties are carboxyl groups [43]. Moreover, according to the result by Dimiev et al. (2012), there is one acidic site for every six to eight carbon atoms; i.e., 100 g of GO sample contains 500-800 mmol of active acidic sites [54]. The adsorption capacity of the present GOs (≈3200 mmol/100 g) ( Table 2) is larger than the concentration of the active acidic site. This is due to the formation of a Cs cluster. It is also generally accepted that the basicity and the size of the molecule to be adsorbed are a significant aspect in the adsorption process [55,56]. In this work, the pH of the solution is around 5-6 when the Cs contacts the GO-POM composite or GO only. Some previous researchers have proved that the pH condition in GO samples influences the adsorption capacity. It can be explained that at pH above 3.9 (pHpzc), the surface area of GO is partially negatively charged and the electrostatic attractions between cations and GO are stronger because of deprotonation reaction [57]. It is also caused by the higher availability of active sites on the GO sheets [58]. The other important evidence of Cs adsorption is also visible in PXRD measurement. The sharp and intense diffraction peaks were reduced by introducing Cs (Figures 4 and S5 (see supporting information)). The broad diffraction of GO c70 , GO c72 , GO c39 , and GO c40 showed the non-uniform surface complexation with the Cs cation. PXRD and TEM show the formation of the Cs cluster. This might correspond to the origin of GO acidity, which contributes to the high cation exchange capacity (CEC). The GO acidity is made by oxygen functional groups, and the main moieties responsible for the acidic properties are carboxyl groups [43]. Moreover, according to the result by Dimiev et al. (2012), there is one acidic site for every six to eight carbon atoms; i.e., 100 g of GO sample contains 500-800 mmol of active acidic sites [54]. The adsorption capacity of the present GOs (≈3200 mmol/100 g) ( Table 2) is larger than the concentration of the active acidic site. This is due to the formation of a Cs cluster. It is also generally accepted that the basicity and the size of the molecule to be adsorbed are a significant aspect in the adsorption process [55,56]. In this work, the pH of the solution is around 5-6 when the Cs contacts the GO-POM composite or GO only. Some previous researchers have proved that the pH condition in GO samples influences the adsorption capacity. It can be explained that at pH above 3.9 (pH pzc ), the surface area of GO is partially negatively charged and the electrostatic attractions between cations and GO are stronger because of deprotonation reaction [57]. It is also caused by the higher availability of active sites on the GO sheets [58].  As discussed above, these data have a good correlation with the information tha observed in Raman spectra. Based on Raman analysis (Figures 5 and S9-S12 (see suppo ing information)), it was found that the change in spectrum by Cs adsorption was sim for both the GO-only samples and composite samples. It can be seen that the second-or peaks have changed after Cs adsorption. In previous studies, it was strengthened that second-order peak indicates that the crystallite size has changed. This indicator show that the spectrum broadens consistently with decreasing the domain size [59]. Therefo during the chemical processing (Cs adsorption), the significant structural change of carbon framework occurred. In addition, based on the sample experiment us  Figure S9, see supporting information). It might be caus by the interaction of oxygen functional groups and the hole defect in the GO sheets w Cs cation [58]. Thus, the change that the crystal structure after Cs adsorption has alter by amorphization could be confirmed. As discussed above, these data have a good correlation with the information that is observed in Raman spectra. Based on Raman analysis ( Figures 5 and S9-S12 (see supporting information)), it was found that the change in spectrum by Cs adsorption was similar for both the GO-only samples and composite samples. It can be seen that the second-order peaks have changed after Cs adsorption. In previous studies, it was strengthened that the second-order peak indicates that the crystallite size has changed. This indicator showed that the spectrum broadens consistently with decreasing the domain size [59]. Therefore, during the chemical processing (Cs adsorption), the significant structural change of the carbon framework occurred. In addition, based on the sample experiment using [GO 39 Figure S9, see supporting information). It might be caused by the interaction of oxygen functional groups and the hole defect in the GO sheets with Cs cation [58]. Thus, the change that the crystal structure after Cs adsorption has altered by amorphization could be confirmed.  Table 3 shows the result of Cs adsorption by using the GO-POM nanocomposite GO-POM nanocomposite (Table 3) increased the Cs adsorption capacity effectively pared to the adsorption capacity of the original material ( Table 2). The change ran around 6.7% to 42.9%. This result is consistent with the previous finding that signified characteristic of surface area, showing an important role of POM species with carbon terials [60,61]. As a consequence, the Cs + adsorption becomes high. We compared th erage increment of Cs adsorption of the composite material between GO having 70-7 % C element and GO having 39-40 wt % C element. The functional groups of GO pro a binding point for chemical modification [25]. It is seen in Figure 6 that there is no si icant difference between GO having 70-72 wt % C element and GO having 39-40 wt element.   Table 3 shows the result of Cs adsorption by using the GO-POM nanocomposite. The GO-POM nanocomposite (Table 3) increased the Cs adsorption capacity effectively compared to the adsorption capacity of the original material ( Table 2). The change range is around 6.7% to 42.9%. This result is consistent with the previous finding that signified the characteristic of surface area, showing an important role of POM species with carbon materials [60,61]. As a consequence, the Cs + adsorption becomes high. We compared the average increment of Cs adsorption of the composite material between GO having 70-72 wt % C element and GO having 39-40 wt % C element. The functional groups of GO provide a binding point for chemical modification [25]. It is seen in Figure 6 that there is no significant difference between GO having 70-72 wt % C element and GO having 39-40 wt % C element.  We compared the average increment of Cs adsorption of the composite material between the concentration ratio 1:8 and concentration ratio 4:1, but we could not find the significant difference. In principle, the adsorption behavior of the GO-POM nanocomposite might be influenced by the significant role of GO as the electron acceptor and Dawsontype POM as the electron donor that led to enhance the attraction between Cs + and the composite [62]. The other phenomena authenticated by another researcher confirmed that the interaction between POMs and GO can provide more mobile protons and flexible pathways [63]. In addition, Wang's group reported that the surface charges of GO modified by functional groups (GO-COOH, GO-NH 2 , GO-OCH 3 ) can be used to modulate the charge transport [64]. Therefore, the presence of Dawson-type POM in the GO system enhanced the negative charge of the GO part in the composite and provided a good adsorption of Cs + ion ( Figure S3, see supporting information). The interaction of Cs + ion with Dawson-type POM was an ion-pair complex [65]. The anion surface of Dawson-type POM is electron rich and has a strong interaction with Cs + . Therefore, the introduction of Dawson-type POM in our adsorbent system was important to enhance the adsorption of the Cs + ion. However, further investigation is still needed to find out the mechanism that occurred behind the phenomenon.

Adsorption Efficiency and Adsorption Capacity of GO-POM Nanocomposite
It is also reported that the mechanism of cation-GO interaction in the adsorption behavior should be considered together with aggregation process [66]. In the present study, Figure 7a,b exhibited that the aggregation occurred after Cs adsorption by using the GO-POM nanocomposite. The results demonstrated that the composite samples immediately coagulate after Cs adsorption except for [GO 70 POM] 41 , which still forms a stable solution after one hour keeping at ambient temperature (see Figure S13, supporting information). We predicted that a small granule of Cs was formed, and there was a crinkled GO sheets coating around it (see Figure S16, supporting information). It can be signified that a stronger bond was formed. It might be predicted that there is a different composition of oxidative debris (OD) component among the GO samples. As a previous study explained, the OD of GO has a significant role in the origin of GO acidity [67]. This is already discussed in the previous section (Section 3.7.1 regarding the contribution of GO acidity to the cation exchange capacity). However, further investigation is still needed to prove it and to obtain strong evidence. Unfortunately, the elemental analysis of Figure S16 (supporting information) was not conducted. Regarding this phenomenon, it can be assumed that there are two important factors that might contribute to the aggregation behavior. First, the size of the GO and composite sample matter. As presented in Table S2, the size of each sample is diverse. The range of values is around 4.3 to 34.78 nm. A previous study showed that the size of GO (nano GO, colloid GO, micron GO) can strongly influence the aggregation behavior [68]. Second, the other factor is the degree of acidity. It is strongly related with the functionalities of the GO sample. It was explained in an earlier section in this work that there are two types of GO samples, which have different C/O compositions. Evidence was shown by the previous study that the GO (single or multilayers) coagulates in highly acidic conditions because of losing their surface charge [54]. It is also correlated well with other work that identified the effect of cations on aggregation [47]. Moreover, it was strengthened in the previous theoretical analysis that considered the important role of H 2 O as a dipole molecule involving cation-π interaction. Therefore, the electrostatic interaction would be significantly influenced [69,70]. As calculated in the previous studies, the value of the internuclear distances d ion-water of Cs + is 0.315 nm [71]. Therefore, the interaction of Cs + with the GO-POM nanocomposite becomes easier to diffuse through the interlayer distance of the GO-POM nanocomposite, which has a d-spacing around 0.69-0.81 nm and 0.77-1.02 nm for the GO sample (see Figure S11, supporting information).
interaction would be significantly influenced [69,70]. As calculated in the previous studies, the value of the internuclear distances dion-water of Cs + is 0.315 nm [71]. Therefore, the interaction of Cs + with the GO-POM nanocomposite becomes easier to diffuse through the interlayer distance of the GO-POM nanocomposite, which has a d-spacing around 0.69-0.81 nm and 0.77-1.02 nm for the GO sample (see Figure S11, supporting information). By FTIR spectra, specific changes are marked at 1150-1350 cm −1 after Cs adsorption in all GO-POM nanocomposites. Especially for composites that consist of ≈39 wt % C element, the IR spectrum shows more significant changes (see Figures 8e-h and S7A in the supporting information), whereas for the composites that consist of 70-72 wt % C element, the peak shifts are not so significant, thus making analysis quite difficult (see Figures 8ad and S7B in the supporting information). A similar tendency also occurred in all GO samples after Cs adsorption (see Figure S6 in supporting information). This result indicates that there is a change in the interlayer chemistry of the sample that influences the defect (hole) formation and functional group formation. It might be caused by the interaction between the GO functional group (including carboxyl, phenol, multi-group ether) and the Cs cation [58]. As explained by another researcher, the pH-driven reversible epoxide formation in GO (epoxy opening/closing reactions) should be considered as an important part of the reactivity and properties of GO [72]. The tendency is observed that the interlayer distance after Cs adsorption in all nanocomposite samples were expanded, even though the increment was not so high (see Table S2 to 7.9 A, and 6.9 to 7.7 A, respectively. In addition, the other study confirmed that the CO2 gas is trapped within the multilayers of GO (marked at 2357-2359 cm −1 ). Intentionally, they verified that the intercalated water has an essential role for accelerating the reaction by some specific mechanism [73]. By FTIR spectra, specific changes are marked at 1150-1350 cm −1 after Cs adsorption in all GO-POM nanocomposites. Especially for composites that consist of ≈39 wt % C element, the IR spectrum shows more significant changes (see Figures 8e-h and S7A in the supporting information), whereas for the composites that consist of 70-72 wt % C element, the peak shifts are not so significant, thus making analysis quite difficult (see Figures 8a-d and S7B in the supporting information). A similar tendency also occurred in all GO samples after Cs adsorption (see Figure S6 in supporting information). This result indicates that there is a change in the interlayer chemistry of the sample that influences the defect (hole) formation and functional group formation. It might be caused by the interaction between the GO functional group (including carboxyl, phenol, multi-group ether) and the Cs cation [58]. As explained by another researcher, the pH-driven reversible epoxide formation in GO (epoxy opening/closing reactions) should be considered as an important part of the reactivity and properties of GO [72]. The tendency is observed that the interlayer distance after Cs adsorption in all nanocomposite samples were expanded, even though the increment was not so high (see Table S2, supporting information). For example, in the [GO 40 18 composites, where the d-spacing was expanded from 7.2 to 7.9 A, 8.1 to 8.6 A, 7.3 to 7.9 A, and 6.9 to 7.7 A, respectively. In addition, the other study confirmed that the CO 2 gas is trapped within the multilayers of GO (marked at 2357-2359 cm −1 ). Intentionally, they verified that the intercalated water has an essential role for accelerating the reaction by some specific mechanism [73].
Another work also confirmed that the functional group of GO can act as a ligand to replace the water molecule from the metal cations in the first coordination sphere. Previous work verified that the relaxation time by using the NMR relaxation method can be considered as an essential variable for identifying the metal complexing mechanism of GO in different charge, size, and electron configuration of the metal as a function of pH and GO concentration [74]. It was also confirmed by another study that the presence of wrinkle-like water tunnels has a strong relation with the water adsorption ability of GO [75]. Another work also confirmed that the functional group of GO can act as a liga replace the water molecule from the metal cations in the first coordination sphere. P ous work verified that the relaxation time by using the NMR relaxation method ca considered as an essential variable for identifying the metal complexing mechanis GO in different charge, size, and electron configuration of the metal as a function o and GO concentration [74]. It was also confirmed by another study that the presen wrinkle-like water tunnels has a strong relation with the water adsorption ability o [75].
Deeper investigation into carbon species in each composite sample was perfor This is shown in Table 4, where the graphitic zone can be calculated by using the grated intensity of graphitic (2θ = 26.6) and oxidized peaks (2θ = 8.  in the P pattern. The amount of graphitic zone (G%, unoxidized region) was calculated u Equation (7), as follows: where Igraphite and Ioxidized are the intensities of the graphitic and oxidized diffraction p of GO in the PXRD pattern, respectively [76]. As presented in Table 4, the diversification of sp 2 /sp 3 carbon clusters on each sa has been produced. This variance among the samples indicates that there is a differ in the degree of oxidation. However, a peculiar phenomenon was seen in the GOc70 ple. The data in Table 4 show that the percentage of the oxidized zone of GOc70 is h  Deeper investigation into carbon species in each composite sample was performed. This is shown in Table 4, where the graphitic zone can be calculated by using the integrated intensity of graphitic (2θ = 26.6) and oxidized peaks (2θ = 8.  in the PXRD pattern. The amount of graphitic zone (G%, unoxidized region) was calculated using Equation (7), as follows: where I graphite and I oxidized are the intensities of the graphitic and oxidized diffraction peaks of GO in the PXRD pattern, respectively [76]. As presented in Table 4, the diversification of sp 2 /sp 3 carbon clusters on each sample has been produced. This variance among the samples indicates that there is a difference in the degree of oxidation. However, a peculiar phenomenon was seen in the GO c70 sample. The data in Table 4 show that the percentage of the oxidized zone of GO c70 is higher compared to those of GO c72 , GO c39 , and GO c40 . The ratio of the oxidized zone is not much different for GO c72 , GO c39 , and GO c40 . The GO c70 shows the shorter interlayer distance than GO c72 , GO c39 , and GO c40 . This may cause the peculiarity of GO c70 . However, the peculiarity of GO c70 is still under study. According to the literature, the ratio of the oxidized zone does not have a linear correlation with the oxidation ratio [77]. The unusual characteristic of GO c70 also has a direct effect on the composite material produced by using GO c70 (as provided at the beginning of Section 3.7.2). Further investigation is needed to obtain more understanding of the dependence of these factors. A recent study revealed that the ratio C/O has a big influence on the physical properties because of the change of the surface speciation [78]. Another investigation also confirmed a strong correlation between the content of Lewis acid sites (via basal plane epoxide) and the Bronsted acidic carboxylic group in the GO sample with a strong acidity and high oxidation degree [79].  41 , which consists of GO c70/ POM with a concentration ratio of 4:1, was decreased. It correlated well with the in-plane crystallite size of each composite material that expanded (see Table S2, supporting information). Particularly, the GO-POM nanocomposite with a concentration ratio of 4:1 shows an in-plane crystallite size larger than the GO-POM nanocomposite with a concentration ratio of 1:8. It could be assumed that there is an electron transfer from POM species to GO [80,81]. Another researcher also confirmed that the surface area of the GO-POM composite was improved five times due to the covalent bond between the POM species and GO [38]. In general, the enhancement of adsorption capacity is in line with the increase of sp 2 /sp 3 carbon clusters on each composite surface. Based on the result in Figure 6, the lowest value of the percentage increase of Cs adsorption capacity is observed in [GO 70 POM] 41 . It is only 6.7%. This might be caused by the decreased oxidized zone (sp 2 /sp 3 hybrid carbon) after the composite was formed (see Table 4). Table 4 shows that the value decreased from 30.3% (GO c70 ) to 28.6% (GO c70 /POM composite). It is also confirmed by the PXRD pattern, which shows the same tendency in [GO 70 POM] 41 : that the diffraction has reduced from 11.57 • to 10.99 • if compared with GO c70 as a precursor material. Some researchers have tried to address the complexity of GO especially to produce a graphene-based hybrid material for diverse applications. For example, one researcher has attempted to control the sp 2 /sp 3 hybrid carbon structure in GO by a simple ethanol solvothermal method [82]. Furthermore, a similar trend was investigated to enlarge the carbon cluster size by increasing the sonication time [83].

Conclusions
In this work, the GO-based POM nanocomposite was successfully obtained. The Cs adsorption capacity increased by forming a nanocomposite. The result proposed, first, that the C/O ratio of the GO sample has significantly influenced the characteristic of the GO-POM nanocomposite for the Cs adsorption performance after forming the composite. Specifically, the GO sample, which consists of 70-72 wt % C element, is a good precursor material to incorporate Dawson-type POM compared to the GO that has ≈39 wt % C element in their atomic composition. Second, the concentration ratio of GO/POM in forming a composite material can be considered to maximize the adsorption performance. Additionally, the size and the acidity of the GO also influence the aggregation state. However, further investigation is needed to know the size tunability and the role of acidic properties of GO in different C/O compositions for Cs adsorption. The further calculation was performed to quantify the graphitic zone in each sample. The result revealed that the oxidized zone was enhanced after forming a composite in all samples except [GO 70 POM] 41 , which shows a different character. Further scrutiny is needed to open the understanding of the role of the oxidative debris (OD) component of GO and the level of hydration of GO for adsorption capacity.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/ma14195577/s1, Figure S1. SEM measurement (a) α-K 6 P 2 W 18 O 62 .14H2O (b) Tungsten element (c) Oxygen element (d) Carbon K-α X-ray element (e) Secondary electron image. Figure S2. SEM measurement, [GO 70 POM] 41 (a) Secondary electron image (b) Carbon K-α X-ray element (c) Oxygen element (d) Tungsten element (e) Aluminum element (f) Cesium element (g) Chloride element. Figure  S3. Plausible Interaction between Graphene-Oxide (GO) and Dawson-type Polyoxometalates (POMs). The black and the grey balls corresponds to carbon and hydrogen. The polyhedral model represents to Dawson-type polyoxometalates. The blue dashes indicated of the presence of interaction between GO-POMs. Figure S4. Xrd pattern of α-K 6 P 2 W 18 O 62 .14H 2 O. Figure S5. XRD pattern of GO c39 , GO c40 , GOPOM composite before and after Cs adsorption. Figure S6. IR spectra for (a) GO c70 (b) GO c72 after Cs adsorption (c) GO c72 (d) GO c72 after Cs adsorption (e) GO c39 (f) GO c39 after Cs adsorption (g) GO c40 (h) GO c40 after Cs adsorption. Figure S7. IR spectra for (a) [GO 39 41 after Cs adsorption. Figure S8. IR spectra for (a)POM (b)POM after Cs adsorption. Figure S9. Raman spectra (a) Graphite, GO c70 , GO c72 , GO c39 , GO c40 , (b) [GO 70 POM] 41 and after Cs adsorption. Figure S10. Raman spectra composite (a) [GO 39 POM] 41 before and after Cs adsorption (b) GO c70 before and after Cs adsorption. Figure S11. Raman spectra (a) GO c39 before and after Cs adsorption (b) GO c40 before and after Cs adsorption. Figure S12. Raman spectra composite [GO 72 POM] 41 before and after Cs adsorption. Figure S13. The composite solution after Cs adsorption. Only Composite [GO 70 POM] 41 (b) that forms stable solution (black color). Another composite solution shows remains coagulated. The photo was taken after more than one-hour stay at ambient temperature. Figure S14. The GO solution after Cs adsorption. All GO solutions shows remains coagulated. (a) GO c70 and (b) GO c72 shows different coagulated behavior with (c) GO c39 and (d) GO c40 . The photo was taken after more than one-hour stay at ambient temperature. Figure S15. Coagulation process after Cs interaction with (a) GO c70 and (b) GO c72 . The photo was taken after more than one-hour stay at ambient temperature. Figure S16. SEM image by using composite (a) [GO 70 POM] 41 after Cs adsorption and (b) Cs particle (zoom mode based on figure a). Table S1. The ratio I D /I G of Graphene oxide and composite before and after Cs adsorption. Table S2. The crystallize size of GO.