Preparation and optimization of novel graphene oxide and adsorption isotherm study of methylene blue

2.1 Materials

Graphite powder (particles size < 50 ϻm, purity >=99.5%), potassium permanganate (KMnO₄), sulfuric acid (H₂SO₄), phosphoric acid (H₃PO₄), hydrochloric acid (HCl), and hydrogen peroxide (H₂O₂) and also methylene blue (C₁₆H₁₈ClN₃S) were purchased from Merck company. All of the materials used were analytical grade. Deionized water (DI) was used in all experiments.

2.2 Equipment and devices

The Ultrasonic bath (PARSONIC 7500 s with Operational voltage 50 Hz, and frequency 28 ± 5%KHz, ultrasonic power100 Watt Max, Iran), Magnetic stirrer, IKA, Shaker (KS 260 basic, company IKA, Germany) and centrifuge (ILettich, ZENTAIFUGEN max.55) are used in the synthesis and adsorption experiments. In order to characterize and determination of the functional groups and study morphology and the structure of samples, the following devices have been used:

UV–visible spectra of the samples were registered using SPECORD 210 PLUS analytic Jena co, Germany. Fourier Transform Infrared (FTIR) transmittance spectra of the GO specimens were operated by TENSOR 27 Bruker CO, the USA, within the range of 400–4000 cm⁻¹. The morphology of the samples was identified by field emission scanning electron microscopy (FE-SEM, model SIGMA VP-500, ZEISS CO, Germany). To study the weight changes of the samples with temperature during the time, Thermogravimetric analysis (TGA, PC Luxx 409,) was conducted from 20 to 800 °C under a nitrogen atmosphere with a heating rate of 10 °C/min. X-ray diffraction (XRD, D500 Siemens, Germany) was used with monochromatic Cu Kα radiation (λ = 1.5406 Å) in the 2ϴ range of 5–80° at ambient temperature.

2.3 Synthesis of graphene oxide (GO)

Graphene oxide was synthesized from graphite powder by altering the improved hummer method (Marcano, et al., 2010). 0.75 gr of graphite was added into the reaction flask, then 90 ml sulfuric acid and 10 ml phosphoric acid were added into the Erlenmeyer while was in an ice bath and mixed by a magnetic stirrer. Followed by 4.5 gr of KMnO₄ was added gradually during 3–7 hr. After that, the resulting solution was placed in the ultrasonic bath for a predetermined time and temperature. The solution was cooled down to room temperature. Then, 120 ml deionized water with 3 ml 30% H₂O₂ was poured. The color of the solution became yellow-brown. Finally, the acquired product was washed with 5% HCl solution followed by washing with deionized water till nuratual pH. The product centrifuged to precipitate and placed in a vacuum oven at 50 °C until it reach constant weight.

In order to the progress of the oxidation degree, optimizing the reaction conditions, reducing the time of reaction and number of experiments, an experimental design with the RSM technique was employed. The advantage of RSM toward the previous methods (Ex: Taguchi) remarks the interactions between parameters, while the other procedures evaluate one parameter's effect whereas the remnants are invariant. Finally, treacherous optimal conditions are achieved (Hosseinpour et al., 2011; Dan et al., 2020). For this purpose, three effective parameters on the synthesis procedure in this method were considered. These parameters are ultrasonic bath temperature, ultrasonic operating time, the time of adding the oxidizing agent. The selected range of these parameters are reported in Table 1.

The total number of experiments was 17, that the conditions and adsorption intensity of the synthesized GO are shown in Table 2.

2.4 Adsorption experiment of methylene blue

To evaluate the adsorption capacity of the synthesized graphene oxide, isotherm of Methylene Blue (MB) adsorption on the graphene oxide synthesized in the optimum conditions (determined with DOE) were investigated. Solutions with a predetermined initial concentration of MB in the range 240–960 ppm was prepared. Then 10 mg of GO was added to the solutions. The solutions with pH 6–7 were placed on the shaker incubator at 150 rpm and room temperature for a different resident times in the range of 5–30 min.

Finally, the adsorbents were separated from the solution, and concentration of the MB solutions were measured by UV–Vis spectrophotometer at maximum absorbance (λ_max = 664 nm). The calibration standard curve was used to measure the concentration of the methylene blue unknown solutions after adsorption. The equilibrium adsorption capacity of the adsorbent $q_e$ (mg/gr) is presented from the following equation:

$c_0$ (mg/L) and $c_e$ (mg/L) are initial and equilibrium concentrations of methylene blue in solution, respectively v (L) is the solution volume and w (gr) is the mass of adsorbent. The removal percent (R%) of MB can be obtained from the following equation:

2.5 Cyclic test and effect of initial pH

Cyclic test has been carried out to study the regenerative capability of the synthesized GO. For this purpose, the GO recovered after the adsorption process using centrifuge and washed with acetic acid solution to remove the MB then followed by deionized water until neutral pH. Then the recovered adsorbent used for another cycle.

As for the effect of pH on the adsorption process, the initial pH of the GO and MB was set to 3–11 using HCl and NaOH solutions. Initial concentration of MB set to 800 ppm, 10 mg GO added to the solution followed by mixing for 10 min to reach the equilibrium state, and Ce was measured.

3.1 Experimental design and optimization method

To determine the effective parameters on the response of the system, the analysis of variance (ANOVA) was used. Some of the parameters do not affect system response, which can be omitted. The parameters influencing system response were determined via p-value by software. The factors with a significant effect, have a p-value less than 0.05. Hence, the parameters of the oxidation degree (with p-value larger than 0.1) were preserved. Owing to this probability has existed that these parameters affected on system response. The p-values related to the beneficial parameters on the oxidation degree model are presented in Table 3.

A, B, C parameters were indicators of ultrasonic temperature, ultrasonic time and the time of adding oxidant, respectively. According to the residual parameters, the quadratic model for optimizing the oxidation degree was reported. Regarding, the model regression R-squared parameter was 0.8585. Due to the p-value of the model (<0.0001) and the model regression R-squared parameter equal to 0.8585, it can be said that the model prediction is valid. Based on the p-value of the temperature and ultrasonic time, these factors were effective on the system's response. P-value of lack of fit was 0.9203, showed that this parameter did not affect the model; And also, the value of adequate precision shows the comparison between the predicted range and the foretold average error by model. If it is more than 4, the model can predict the data correctly. Here the value of adequate precision is 16.068. The proposed model for oxidation degree that only the effective parameters on it were considered, is presented through Eq. (2). In this equation, the individual effects, interactions of parameters and also the effects related to their curvature have been supposed.

3.2 Model accuracy control

In this section, the accuracy of the proposed model using different statistics charts has studied. The normal function diagram of the remaining parameters has been shown in Fig. 1. As all of the data were around the line and followed the linear pattern that they didn’t indicate to the abnormal error term (Dan et al., 2020). The comparison of the values estimated by the model versus experimental values has been shown in Fig. 2. The data were around the line (as shown in Fig. 1) the predicted model was matched with experimental data. Whatever the data were closer to the line, the proposed model had more consistent with the experimental data which reported in Table 2.

Close

Fig. 1

Normal function of the remaining parameters.

Export to PPT

Close

Fig. 2

Predicted value versus the experimental data.

Export to PPT

3.3 Model prediction

One of the best ways to understand the effect of the parameters is the 3D chart. In this case, the effect of temperature (parameter A) and ultrasonic time (parameter B) are shown in Fig. 3. These two parameters within the range of the investigation were effective on oxidation degree, while the time of adding oxidant (parameter C) did not affect the studied range.

Close

Fig. 3

Three dimentilnal response surface chart of process varriable vs oxidation degree.

Export to PPT

Due to the investigated temperature range 25–55 °C according to the RSM, in the high temperatures, there was a possibility of excessive graphite oxidation. With increasing the temperature during the oxidation process, carbon in the form of CO₂ from the carbon structure was exited, which caused the defects in the structure of graphene oxide. In addition, in high temperatures, due to the high surface energy, the particles tended to agglomerate, which was the reverse phenomenon that led to a decrease in the oxidation degree. On the other hand, in high temperatures, there was the probability of decomposition of the hexagonal structure of graphene oxide.

Penetration of the oxidizing agent into the graphite layers was happened in a slow process and needed to assign enough time. Ultrasonic, due to the generation of intense sound cavitation event which produce the bubbles at the points of contact of liquid and surface of graphite, exfoliation of graphite occurrs. As a result, it accelerated the process. The adding time of the oxidant had no main effect on oxidation degree in the range of study (3-7hr (and ultrasonic did the penetration process of oxidizing agents between layers. So, if it was not considered the time at the beginning of the reaction in order to add the KMnO₄, in the final product there might have been a lot of graphite, which led to a decrease of oxidation degree.

The effect of ultrasonic mechanism mentioned above, that contains generation, progress, exploding of bubbles, high temperatures and pressure were created which expedited the peeling (Safarifard and Morsali, 2012) and the time of exfoliation was reduced. On the other hand, peeling took place in two forms: simple and rigid. Facile exfoliation is done in two step first step is on the surface and the second step is on the depper layers of the compound.

So, the graphite skeleton in a gradual process has been oxidized which caused the deformation in its structure. Also, it made the oxidation degree to be distinct. As a result, it can be concluded that the oxidized graphite had a heterogeneous structure with profound oxidation on the edges and low oxidation in the inner regions. Thus, it can be said the variation in the intensity of peeling could have made a discrepancy in oxidation degree.

3.4 Determination of the optimal conditions

As previously mentioned, to progress the oxidation degree, the RSM method has been used. Based on that, adsorption intensity in 300 nm and operational variables were in the selected range. The predicted optimum conditions have been reported in Table 4. The results of the experiments showed good compatibility with the predicted model. Based on the predicted model, the optimum conditions of synthesis in order to maximize the oxidation degree was: ultrasonic temperature: 30.8 °C, ultrasonic time 47.5 min and adding time of the oxidant equal to 3 hr.

3.5 The characterization of the synthesized graphene oxide

The optimal condition for the graphene oxide with the maximum oxidation degree obtained from design of the experiment. This product was synthesized and then characterized with several analyses such as UV–Vis spectroscopy, Fourier Transform Infrared, Thermogravimetric analysis, X-ray diffraction and Field Emission Scanning Electron Microscopy. The results are reported in the following sections.

3.5.1

3.5.1 UV–vis spectroscopy

To make a uniform solution of graphene oxide for accomplishing UV–Vis analysis, 12.5 mg of graphene oxide was weighed and dispersed in 100 ml of Dionized water. The UV spectrum of the graphene oxide is presented in Fig. 4. Carbon substances contain a wide peak related to sp2 feature at the range of 200–250 nm in the UV spectrum (Russo et al., 2012). Graphite has a peak with the low intensity in 275 nm (Senthil et al., 2017). Graphene oxide has two peaks, the first peak at 228 nm is related to π-π* transition of the C—C aromatic ring bonds and the second one is in a form of shoulder around 300 nm which indicate n-π* carbonyl group bonds (Jasim et al., 2016; Yu et al., 2016) and represented the intensity of the oxidation degree (OD). The absorption peak in 228 nm is the characteristic absorption band and used to identify the synthesized graphene oxide from graphite OD can be achieved via max λ of UV spectrum (Marcano, et al., 2010). Comparing the results of UV–VIS λ max wave length data and comparing with other traditional method Hummer’s method, improved Hummer’s method and Modified Hummer method (Table 5) indicate that the use of ultrasonic has positive effect on the synthesis of GO.

Close

Fig 4

The diagram of UV–Vis of graphene oxide spectrum.

Export to PPT

Table 5 λ max Comparison between different synthesize of GO.

Component	λ max(nm)	References
GO	238	(Hidayah, Oct. 2017)
improved Hummer’s method	232	(Mahto et al., 2016)
Modified Hummer method	232	(Mahto et al., 2016)
GO	228	This work

As π-π* Transition increases, it results in lowering the energy, which led to higher λmax, and greater absorption rate in 300 nm, exhibited a higher oxidation degree. Absorption intensity at 300 nm for the optimized graphene oxide is equal to 2.11 which exebits a little error related to the predicted data by the model. Moving the π-π* adsorption peak into a lower wavelength in effect of growing the sp3 expressed higher oxidation degree due to the sonic cavitation occurrence of ultrasonic, more oxygen groups and the radicals have been created between carbon layers. The position of the π-π* bond in graphene oxide has been shifted to lower wavelengths due to the presence of oxygen groups in graphene oxide.

3.5.2

3.5.2 Fourier transform infrared

FTIR analysis was carried out to identify the functional groups on the surface of graphene oxide. The FTIR spectrum of graphene oxide shown in Fig. 5 a demonstrates the presence of the O—H characteristic peak within the range of 3000–3600 cm⁻¹. The absorption peak in 1731 cm⁻¹ is attributed to C⚌O and the absorption band at 1621 cm⁻¹ can be related to C⚌C stretching vibrations of the water molecule between layers. The stretching vibrations of absorbed water molecules appear at 3425.18 cm⁻¹. C—O in epoxy and C—O in carboxyl were located in 1061.56 and 1225.68 cm⁻¹, respectively (Guo et al., 2009; Salehi et al., 2017). Absobrtion band at 1417.03 cm⁻¹ related to O—H deformation. The existence of the oxygen groups in the spectrum of graphene oxide displayed good oxidation of graphite. Spectra b in Fig. 5 represent the GO after the adsorption process as can be seen in the all the characteristic adsorption band are GO exist and new adsorption bands related to the Methylene blue structure added to the spectra which proves the adsorption of methylene blue on the GO.

Close

Fig. 5

The FT-IR spectra of synthesized graphene oxide and after the adsorption process.

Export to PPT

3.5.3

3.5.3 Thermal gravimetric analysis (TGA)

Thermal degradation of graphite and graphene oxide by analysis of TGA and DTA has been investigated. Graphite is thermally stable at temperatures lower than 600 °C, and it exhibits a stage of weight loss at a temperature above 600 °C (Hu, 2011). The presence of oxygen functional groups on graphene oxide reduce the thermal stability of the graphene oxide (Yang et al., 2009). According to Fig. 6 TGA and DTA data, graphene oxide degraded within three steps. The first weight loss step was 3.80%, at the temperature lower than 120 °C related to the evaporation of water molecules, which was due to the hydrophilic nature of graphene oxide and moisture. The second stage of weight loss was about 29.75% within the range of 120–240 °C which attributed mainly to the decomposition of the hydroxyl, epoxy, carboxylic acid functional group on the surface. Another 6.86% weight loss was within the range of 240–500 °C, due to eliminating the remaining interlayer water molecules and sulfate groups. The disintegration of the unstable oxygen groups led to production of CO, CO₂, and steam (Stankovich, 2007; Kuila et al., 2012) which caused rapid thermal expansion of the compound. The next stage of the thermal degradation weight loss was equal to 41.93%. graphene oxide at temperatures higher than 500 °C undergo decomposition of stable oxygen functional groups (phenol and carbonyl) and ustable carbon residues in the structure, also oxidation of carbon and burning of the carbon skeleton and pyrolysis of the structural sheets. Total weight loss of graphene oxide in the thermal gravimetric analysis till 800 °C was equal to 83.35%.

Close

Fig. 6

TGA plot of graphene oxide.

Export to PPT

3.5.4

3.5.4 Field emission scanning electron microscopy and Brunauer–Emmett–Teller analysis

To study the surface morphology and surface area of the synthesized compound, FE-SEM and BET analysis was performed. The sheets in the structure of graphite accumulated on each other of via the weak Van der Waals forces. Using ultrasonic deform the structure of graphite sheets and effectively peeling them of into the graphene oxide layers. The expanded layers in graphite undergo an erosion owing to the resulting micro jets from the bubbles eruption (Chen et al., 2011). In Fig. 7 (a-b) represent the graphene oxide at different magnifications due to the deformation during peeling process and agglomeration in the drying process graphene oxide sheets has been folded (Pang et al., 2011). The surface area of the synthesized compound determined using BET analysis equal to 315 m²/g. surface morphology and surface area indicated that the peeling of carried out completely successful and a few layer GO synthesized.

Close

Fig. 7

The images of graphene oxide with magnification of 300 (a), 6000 (b).

Export to PPT

3.5.5

3.5.5 XRD analysis

One of the strongest and the most reliable non-destructive techniques for qualitative analysis of the crystalline materials in the form of powder or solid is XRD. Graphite shows a sharp peak about 2ϴ = 26° which corresponding to the interlayer distance (d-spacing) of 0.34 nm due to the regular arrangement in its structure. Fig. 8 represents the xrd spectra of synthesized GO. The characteristic peak (0 0 2) of graphite completely was vanished and the new peak appeared in 2ϴ = 11.147° which corresponded to (0 0 1) diffraction peak of graphene oxide due to the increment of the interlayer distance. Interlayer distance in the GO increased from 0.34 to 0.81 nm due to the addition of oxygen functional groups such as hydroxyl, epoxy, carboxyl, carbonyl that causes the layer to expand and expholiate (Mallakpour et al., 2015). Another peak at 2ϴ = 22° is the characteristic peak of the reduced graphene oxide, that indicated some graphene oxide has been reduced through the synthesize process.

Close

Fig. 8

X-ray diffraction pattern of graphene oxide.

Export to PPT

3.6 Methylene blue adsorption results on the synthesized graphene oxide

In order to study the adsorption capacity of optimized graphene oxide methylene blue has been used. Effect of methylene blue initial concentration and equilibrium time was investigated.

3.6.1

3.6.1 The adsorption equilibrium time of MB

Fig. 9 represent the adsorption of MB on GO at different times intervals at room temperature, as can be seen, adsorption process reach equilibrium in less than 5 min. This rapid adsorption on the adsorbent surface can be pointed to the lack of interior resistor and ability of graphene oxide for dissolution that showed a strong tendency between MB molecules and GO surfaces. By increasing the contact time, as the available sites have been occupied, the dye molecules percolated slowly into the absorbent structure.

Close

Fig. 9

The impact of contact time on the adsorption of methylene blue.

Export to PPT

3.6.2

3.6.2 The isotherm adsorption study of MB on the synthesized GO

the initial concentration of methylene blue solutions studied to determine the adsorption capacity GO. As the initial concentration of MB increased, the adsorption increased gradually until reached the equilibrium and set to level of as shown in Fig. 10. The highest adsorption capacity is equal to 1635.5 mg/gr. Adsorption capacity of the synthesized GO compared with the other adsorbent used for MB, Results shown in Table 6 indicates that the synthesized adsorbent has better performance than other adsorbent.

Close

Fig. 10

Effect of initial concentration of methylene blue solution on A) the its equilibrium on GO B) Removal percentage.

Export to PPT

Table 6 Comparison of maximum adsorption capacities of graphene – based different adsorbents for MB adsorption.

Adsorbent	Adsorption capacity (mg g−1)	Reference
EGO	17.3	(Ramesha et al., 2011)
GO/Co3O4	40	(Pourzare et al., 2017)
GO	243.9	(Li, 2013)
PT-GO	256.58	(Wang et al., 2020)
GO	286.9	(Sabzevari et al., 2018)
GO	331.87	(Xu et al., 2019)
Graphite oxide	351	(Bradder et al., 2011)
3D GO sponge	397	(Liu et al., 2012)
Fe3O4/GO@MF	418.41	(Chen et al., 2019)
RGO	500	(Yang, 2013)
GO	598.8	(Yan, 2014)
PMMA-rGO	698.51	(Mercante, 2017)
GO	700	(Chia et al., 2013)
GO	714	(Yang et al., 2011)
Graphene	1520	(Wu et al., 2011)
Graphene	1523.85	(Liu, 2012)
GO	1635.5	This work

By increasing the initial concentration of the dye, the diffusion intensity of the dye onto the adsorbent increased due to the enhancement of concentration gradient. At low concentrations, due to the presence of abundant adsorption sites, methylene blue molecules creates a sturdy interactions GO sheets. The removal percentage exhibit a decreasing trend starting from 95.5% and decreasing while the initial concentration increase this trend is shown in the Fig. 10B.

3.7 3 Cyclic adsorption and effect of pH

Cyclic adsorption used to determine the regenerative capacity of the synthesized graphene oxide, as shown in Fig. 11 A, the removal percentage of MB reduces due to the loss of the active cite on each cycle, some of the adsorbed MB entrap in the structure of the synthesized GO and reduce the removal percentage.

Close

Fig. 11

A) Removal experiment of MB using recycled GO B) effect of pH on MB removal percentage.

Export to PPT

pH is one of the most important factors in the solid liquid adsorption since pH affects the charge transfer of the components. The initial pH varied from 3 to 11 by adding hydrochloric acid or sodium hydroxide solutions, results shown in Fig. 11B. It was found that the by increasing the pH adsorption of MB on GO increases. This can be described using acid base equilibrium isotherm that at lower pH GO exist in the form of GO-H and MB competes with hydrogel ions, and as pH increase and the GO exists in the form of GO⁻ and adsorption of MB increase.

3.8 Isothermal adsorption equilibrium study

In this section, the adsorption isotherm model that was more consistent with experimental data was selected. The linear form of the Langmuir isotherm equation is presented below:

The slope and intercept were achieved from plotting the Ce/qe vs. Ce chart, that caused the erect line. According the equation $q_{\mathit{max}}$ and $K_L$ can be evaluated. Where the equilibrium adsorption capacity is $q_e$ (mg/gr) and $q_{\mathit{max}}$ (mg/gr) is maximum adsorption capacity that expresses the mg of the adsorbed MB per gr of adsorbent, ${\mathrm{C}}_{\mathrm{e}}$ (mg/L) is equilibrium concentration, $K_L$ is Langmuir constant (L/mg). Another dimensionless parameter is the equilibrium parameter R_L that calculated using the equation below:

$K_L$ (L/mg) is the constant of Langmuir and $C_0$ (mg/L) demonstrates the maximum initial concentration of dye. Where $R_L$  = 1, isotherm is linear, 0 < $R_L$  < 1 shows the desirable adsorption,   $R_L$ > 1 is representive an of undesirable adsorption conditions, $R_L$ = 0 expresses irreversible adsorption (Aghdam et al., 2016).

According to the linear regression of data with equation (3), and high correlation coefficient, the adsorption data showed good agreement with Langmuir isotherm. The slope and intercept of this model were obtained through the constants of the model. The values of the correlation coefficient and the constants of the model are given in Table 7.

According to the obtained value of $R_L$ , since it was between zero and one, it showed the favorable adsorption.

To verify the compatibility of experimental data with the Freundlich model the linear form of model has been used.

The equilibrium adsorption capacity is $q_e$ (mg/gr), Ferrandich constant, which depends on the adsorption capacity (( $\frac{\mathrm{m}\mathrm{g}}{\mathrm{g}\mathrm{r}}){(\frac{\mathrm{L}}{\mathrm{m}\mathrm{g}})}^{\frac{1}{\mathrm{n}}})),\frac{1}{n}$ the adsorption intensity parameter shows the desirable state of adsorption processes. The values of $K_F$ and $\frac{1}{n}$ were obtained by plotting ln ${\mathrm{q}}_{\mathrm{e}}$ vs. ln ${\mathrm{C}}_{\mathrm{e}}$ . The value of $\frac{1}{n}$ is varied between 0 and 1. Whatever, for heterogeneous surfaces, $\frac{1}{n}$ is closer to zero which indicates the n is large and adsorption deviates from linear isotherm. The n values within the range of 2–10 showed strong adsorption and the range of 1–2 displayed relatively hard adsorption. The values of n lower 1, expressed poor adsorption process (Yang et al., 2011). Based on linear regression of experimental data and Eq. (6), the obtained constants are given in Table 8 that shows regard to the calculated correlation coefficient, adsorption data did not fit well with Freundlich isotherm.