Abstract
Green synthesis of graphene from graphite was carried out in molten salt mixture of LiCl-KCl, which is an environmentally friendly alternative to traditional methods using a large number of chemicals. The effect of carbonization temperature on the final properties of graphene products was investigated at a temperature range of 500-800°C. The layer numbers of graphene products were determined as a single layer and few layers by XRD and Raman analyses. The highest conductivity value of 1219 S/m was reached with the sample synthesized at 600°C (GGr600), which is 10-fold higher than that of commercial graphene (CG) (115 S/m). The percentage of carbon contents of GGr600 and sample synthesized at 800°C (GGr800) was determined as 86% and 73% by XPS. From the SEM analyses, it was observed that graphene products have sheet-like structures. A TEM image showed that GGr800 has the thickness changing from 5-9 nm, pointing out the graphene with few layers. AFM analysis demonstrated that GGr600 has a single-layer sheet-like structure.
1. Introduction
Graphene, which is a quasi-two-dimensional carbon, has become a very popular material in recent years with a wide range of applications [1]. Graphene is a single atomic layer of graphite with a 0.34 nm thickness, first discovered by Geim and Novoselov in 2004 [2–4]. It is a sheet of sp2-bound carbon atoms organized into a hexagonal lattice [5, 6]. Graphene has many superior physical and chemical properties such as mechanical strength, thermal conductivity, high mobility, optical transparency, and room temperature quantum Hall effect and excellent electronic properties such as Dirac particles with linear dispersion, transport energy gap, and simply absorption coefficient of lights [7, 8]. In particular, due to its high electrical conductivity, it has many applications such as fuel cells, solar cells, anticorrosive materials, circuit boards, OLED display panels, longer-lasting batteries, and supercapacitors [9–11].
Graphene synthesis methods are divided into two categories such as top-down and bottom-up approaches [12, 13]. In the top-down process, graphitic microstructures such as graphite, carbon fibers, and carbon nanotubes are the starting materials and individual graphene layers are extracted or peeled either by physical, electrochemical, or chemical methods such as the chemical reduction of graphite oxide, among which Hummer’s method is the most popular [14–16]. The bottom-up approach implements carbon atoms as building blocks; epitaxial growth of graphite on SiC surfaces and chemical vapor deposition are the most sophisticated methods [17]. Further, various inexpensive and simple molecules such as naphthalene, benzene, hexane, alfalfa plant, glucose, and cellulose can be utilized as a starting material in this approach [18–22].
Furthermore, various popular techniques such as Hummers’ method, liquid-phase exfoliation of graphite, and chemical vapor deposition suffer from disadvantages such as low quality of the graphene product, low rate of production, and use of hazardous chemicals [23, 24]. Most of the chemical methods, which have been used up to now, include harsh oxidizers such as H2SO4/KMnO4 or carboxylic acid, formic acid, and organic solvents, all of which are environmentally detrimental [25, 26]. The electrochemical exfoliation of graphite into graphene sheets via molten salts is a promising alternative method, which decreases the number of defects in the graphene structure in comparison to traditional chemical-rich methods [27, 28]. The molten salt process is based on the interaction of alkali ions with graphite that enables the formation of graphene nanosheets and flakes [29].
The eco-friendly production of graphene by the electrochemical exfoliation of graphite in molten LiCl (Green Chemistry 18 (2016), 1952-1964), molten NaCl (Journal of Industrial and Engineering Chemistry 52 (2017), 18-27), and molten LiCl-NaCl (Applied Surface Science 490 (2019), 383-394) has been demonstrated [30–32]. Also, the chemical exfoliation of graphite powder in molten LiCl was reported. The molten salt approach has several advantages such as simplicity, high productivity, low cost, and short processing times compared to other methods used for the production of graphene. In this study, the eco-friendly synthesis of graphene in the eutectic salt mixture of LiCl/KCl is reported. It was shown that graphite can be converted into nanostructured carbons in an ionic molten salt (MS) medium, and the process yields purer graphene via tuning of the synthesis conditions. Therefore, a series of experiments at different temperatures were conducted to investigate the effect of temperature on the product quality.
2. Materials and Methods
2.1. Materials
As commercial graphite (CGr), natural flake graphite, grade 3061, was purchased from Asbury Graphite Mills, Inc., New Jersey. Commercial graphene (CG) was obtained from XG Sciences, Michigan, US. Other chemicals used in the experiments were of analytical grade: lithium chloride (Merck), potassium chloride (Riedel-de Haen), sulphuric acid (95-97%, Merck), potassium permanganate (Merck), sodium hydroxide (Sigma-Aldrich), L-ascorbic acid (Carlo Erba), and hydrochloric acid (37%, J.T.Baker).
2.2. Experimental System and Synthesis Method
Graphene was synthesized by heating the graphite under flowing argon or nitrogen atmosphere in a stainless steel reactor surrounded by a ceramic heater. The experimental system is shown in Figure 1. The temperature of the system was controlled by an Elimko E-200 Series digital temperature controller. The graphite was mixed with the eutectic composition of LiCl/KCl (45/55 by weight) with a ratio (1 : 10), and the powders were homogenized by pestling into an agate mortar. The mixture was then filled into a ceramic crucible and loaded into the reactor. First, the system was vacuumed and then flushed with the reaction gas (Ar or N2) for 5 min. After the starting step, the system was heated to the desired temperature with a 20°C min−1 heating rate and kept at this temperature for 5 h. Finally, the system was cooled to room temperature by turning off the power; meanwhile, the gas flow was fed until the temperature fell below 50°C. The acquired block of products was crushed into particles and washed with a sufficient amount of water to remove the salts. The synthesized material was dried overnight in a drying oven at 60°C.

For comparison, reduced graphene oxide (RGO) was synthesized by modified Hummer’s method; 1 g of graphite was added to 50 mL concentrated sulphuric acid while stirring in an ice water bath. Then, 3 g potassium permanganate was slowly added by controlling the temperature under 55°C. After that, the suspension was stirred at 25°C for 25 min followed by 5 min sonication in an ultrasonic bath (Elmasonic S, 30H). After repeating the stirring-sonication step 12 times, the reaction was cooled by the addition of 200 mL distilled water. An extra 2 h ultrasonic treatment was performed. After adjusting the pH to 6 by the addition of 1 M sodium hydroxide solution, the suspension was further sonicated for 1 h. 10 g L ascorbic acid was dissolved in 100 mL distilled water, and then, it was slowly added to the exfoliated graphite oxide suspension at room temperature. The reduction was performed at 95°C for 1 h. The resultant black precipitates were filtered using a cellulose filter paper and washed with 1 M hydrochloric acid solution and distilled water to neutral pH. Finally, the filtrate was dried in a drying oven to obtain the RGO powder.
2.3. Characterization of Synthesized Samples
X-ray diffraction (XRD) patterns were obtained with a Rigaku D-Max 2200 Series equipped with Cu-Kα radiation () at a scanning rate of 3° per minute. The tube voltage was 40 kV, and the current was 40 mA. XPS measurements were made using a Thermo Scientific K-Alpha X-ray Photoelectron Spectrometer, which has energy range between 100 and 4000 eV, a source-defined analysis area from 30 to 400 μm, a 180° double-focusing hemispherical analyser, and a 128-channel detector.
The microstructures of the synthesized graphene were determined using a scanning electron microscope (FEI Philips XL30 SFEG SEM) and a transmission electron microscope (Hitachi HT7800 TEM operating at 120 kV). The atomic force microscope (AFM) images for the graphene samples were taken with a Veeco NanoScope IV at a contact mode.
Raman spectra were recorded in a Renishaw inVia Raman microscope using a 5x optical lens and a 532 nm laser diode with 50 mW as the excitation source. UV-vis spectral measurements were acquired using a PerkinElmer Precisely Lambda 35 UV-vis Spectrometer. An ultrasonicated suspension of the product in water was used for this analysis. The ultrasonication was performed using an ultrasonic (US) generator (Bandelin® HD 2200 SONOPULS, 200 W, 20 kHz) equipped with a horn-type probe which was used to deliver pulsed ultrasound with controllable power.
Particle sizes of the samples were measured using a Malvern Zetasizer Nano ZS Laser Particle Size Distribution Meter. Samples were dispersed in distilled water and were then agitated at 1600 rpm by using an IKA® MS 1 shaker.
Electrical conductivity tests were conducted using a Keithley 2400 SourceMeter. First, synthesized graphene samples and commercial products were placed in a cylindrical copper container, which has a copper cap. Then, graphene powders were compressed by a hydraulic press under 5 MPa for 0.5 h. The electrical resistivities of the graphene powders were determined by a four-point probe method. Pressure was applied by a joiner’s clamp onto the graphene powder in the copper mold during the electrical conductivity measurement.
3. Results and Discussion
The green synthesis of graphene from graphite was conducted by using molten salt mixture LiCl/KCl at different reaction temperatures of 500°C, 600°C, 700°C, and 800°C in order to examine the effect of temperature on product quality.
XRD patterns of all synthesized graphene products as shown in Figure 2 demonstrated a peak at which is characteristic of carbonaceous materials. Although graphene samples and CGr have the same peak position at , the peak intensities of the graphene products synthesized at 500-800°C are lower than those of CGr. This result indicates the transformation of graphite to graphene, which is in accordance with literature [33–35].

The structural analyses of graphene products were performed by using XRD. The graphene thickness can be calculated using Scherrer’s equation, which is expressed by , where is the thickness of the crystallite (thickness of graphene), is a constant dependent on the crystallite shape (0.89), is the X-ray wavelength (0.15406 nm), is the full width at half maximum (FWHM), and is the scattering angle [36, 37]. The number of graphene layers was calculated by substituting the value of obtained from Scherrer’s equation into the following equation: , where is the interlayer space for the (002) peak at [38, 39].
As seen in Table 1, the layer number of CGr was calculated as 10, while the layer number of the CG was determined as 2. The graphene products synthesized at 500, 600, 700, and 800°C have 2, 4, 2, and 3 layers, respectively, which confirms the obtained graphene structure.
The structure of the graphene products was also evaluated by UV analyses of CG, CGr, GGr500, GGr600, GGr700, and GGr800 given in Figure 3. As seen in the spectrum, characteristic absorbance peaks at around 200 and 280 nm were observed for all samples, which is consistent with the previous studies [39–41].

Figure 4 illustrates the particle size analyses by the dynamic light scattering technique. As seen in Table 2, the particle size values of CGr, CG, GGr500, GGr600, GGr700, and GGr800 are 4180, 2402, 2660, 4205, 3017, and 3459 nm, respectively. When CGr was treated with molten salt mixture of LiCl and KCl, the particle sizes of graphene products were close to the particle size of CG. When the effect of temperature on the extent of exfoliation and reduction of GO was studied, it was observed that at higher temperatures, the crystal size of graphene became closer to that of the graphite crystal [42]. Analogously, in this study, when the reaction temperature increased, the particle diameters of graphene products became larger. On the other hand, the particle size of RGO synthesized by Hummer’s method was found to be 738 nm, which agrees with the literature [43]. The reason for obtaining a smaller particle size in Hummer’s method compared to the molten salt method results from the reaction steps including harsh chemicals and mechanical treatments.

The conductivity values () were estimated according to . In this equation, is the electrical resistance, is the length, and is the cross-sectional area of the sample material. According to Table 2, the electrical conductivity values of GGr samples (GGr500, GGr600, GGr700, and GGr800) are higher than those of CG, CGr, and RGO. Hence, it can be inferred that the molten salt method has an enhancing effect on the conductivity. The increase in the conductivity is due to the advantages of the molten salt method, which enables the production of graphene with high crystallinity and purity [44]. The highest conductivity value was achieved with GGr600 at 600°C. Although the conductivity values belonging to the samples of GGr700 and GGr800 were smaller than those of the samples of GGr700 and GGr800, they are still higher than 115 S/m, which is the conductivity value of CG.
Figure 5 shows the typical Raman spectra of the CG, RGO, and GGr samples obtained at different temperatures. Characteristic graphene peaks were observed at 1345, 1576, and 2659 cm−1 arising from D, G, and 2D bands, respectively. For RGO, GGr500, and CG, the D band becomes stronger and broader because of the higher level of disorders of the graphene layers. GGr700 and GGr800 showed no peaks at the D band, which implies the formation of well-ordered pristine single-layer graphene without a disordered structure [45–47].

In the Raman spectra (Figure 5), the chemical change of graphite to graphene results in a notable decrease in the ID/IG ratio (Table 3). The increase in a broad 2D peak located at around 2600 cm−1 indicates the formation of sp2 carbon atoms at higher heating temperatures [19, 48, 49].
In order to clarify the structure of the graphene products, the ID/IG ratios giving the information on the structural quality and order were calculated. Since the ID/IG ratio of CG is lower than 0.33, the fewer defects and functional groups were observed in the structure. In addition, synthesized GGr700 and GGr800 samples have a smaller D band, which reflects good quality and an ordered graphene structure. The most prominent features in the Raman spectra of the monolayer graphene are the G band appearing at 1582 cm−1 [50]. In fact, the G band is the only band from a normal first-order Raman scattering process in graphene [51]. From the G band, information such as the existence of strain, the effect of gate doping, and the impurity can be obtained about the sample. Since the peak position of the G band is sensitive to the gate voltage applied to the graphene, the Fermi energy can be evaluated by the peak position [52]. The G band arises from the stretching of the C-C bond in the graphitic materials, and it is common to all sp2 carbon systems. The G band is highly sensitive to strain effects in the sp2 system, and thus, it can be used to probe modification on the flat surface of graphene. Therefore, it can be understood that GGr700 has the highest G band, which indicates the high amount of sp2 carbon content.
It is known that if the ID/IG ratio is equal to 0.6, graphene includes few layers. Moreover, when the ID/IG ratio has a value of 0.4, graphene contains a single-layer structure [53]. According to the results given in Table 3, when the higher reaction temperatures were applied in the synthesis of graphene products, the ID/IG ratios decreased. Besides, the ID/IG ratios of CG and RGO are 0.18 and 1.04, respectively. It can be concluded that GGr600, GGr700, and GGr800 have a single layer, GGr500 has few layers, and RGO has multilayers. Therefore, it may be inferred that the molten salt method applied at high temperatures results in graphene with a defect-free structure and single layer.
The presence of sp2- and sp3-bonded carbon atoms and the remaining functional groups were quantified by X-ray photoelectron spectroscopy (XPS; Figure 6, Table 4). The characteristic peaks of C=C, C-O-C, O-C=O, and C-C bonds are located at 284.30, 286.26, 290.34, and 285.11 eV for the GGr600 sample, whereas these bonds arose at 284.24, 286.28, 290.37, and 285.02 eV for the GGr800 sample, respectively. These results are compatible with the previous studies [19, 20, 42, 54].

(a)

(b)
The carbon contents of graphene products are listed in Table 4. As evident from the results, the carbon content decreases by increasing temperature. As the reaction temperature of the carbonization process increases, the carbon fraction of the graphene products generally increases [19]. However, the carbon content of GGr600 is lower than that of GGr800. The oxygen atoms, which originate from the usage of water in the purification steps, may interact with the carbon in the reaction medium. This induced an increase in the amount of functional groups in the structure of GGr800 and caused the graphitic carbon percentage to decrease.
The atomic ratios of C1s and O1s regions belonging to GGr600 and GGr800 samples are given in Table 5. The existence of oxygen content may arise from the treatments after the reaction such as washing with water to remove molten salt from the reaction mixture and storage conditions such as exposure to humid air.
The morphological structures of graphene products were examined by SEM analysis (Figure 7). SEM images of all products demonstrate sheet-like structures such as flakes, edges, and layers. From the image of the GGr600 sample, these sheet-like structures can be seen clearly in comparison to other samples. On the other hand, the SEM images of GGr500, GGr700, and GGr800 samples include the overlapping layers of graphene flakes with arbitrary shapes and random in-plane orientation [19, 20, 54].

These flakes have different sizes and shapes. Edges and intersecting lines of a flake are drawn, and the intersecting lines are used to determine the average flake size. The average size of the flakes () observed in SEM images was calculated as 2 μm by using the following equation [55]:
The sheet-like structure is further confirmed by transmission electron microscopy (TEM) imaging as seen in Figure 8. TEM images display that the graphene structure with different thickness, ranging from 5 nm to 9 nm, indicates the presence of few-layer graphene. Thicker layers are observed in significant amounts indicating that most of the graphene samples have few and multilayers. This deviation from the ideal thickness of 0.335 nm (interlayer distance in graphite along the axis) for chemically derived single-layer graphene can be explained by the presence of a structural disorder and the remaining attached functional groups [19].

The AFM characterization of the final graphene products (GGr500, GGr600, GGr700, and GGr800) was performed to measure the surface roughness and thickness to determine the optimal temperature for graphene growth. Roughness average (Ra), calculated by measuring height alterations on the surface, is less sensitive to big peaks and valleys. The root mean square roughness (Rq) value is the square root of the total of the squares of the particular vertical deviations from the average level [1]. The roughness mean square (RMS) value, which gives the average height deflections of the average level, is estimated by measuring the heights of surfaces of microscopic peaks and valleys. The number of graphene layers, , is calculated via AFM by the following equation assuming the value of single-layer graphene (SLG) thickness as 0.335 nm: where is the thickness determined by AFM to exclude the increase in thickness contributed by the substrate-graphene and graphene-tip interactions. The value of 0.4 is the arbitrary magnitude (equivalent to three graphene layers with 0.335 nm spacing). can be calculated via AFM by measuring the height of the deposited graphene flakes [56]. The height of SLG is strongly affected by the structure. For example, when SiO2 and mica are chosen as substrates, the heights of SLG are ~1 nm and 0.4 nm, respectively [57].
The AFM characterization results of samples (GGr500, GGr600, GGr700, and GGr800) are summarized in Figure 9. The Ra values of GGr500, GGr600, GGr700, and GGr800 are 13.052, 30.110, 17.431, and 16.982 nm, respectively.

The Rq values of GGr500, GGr600, GGr700, and GGr800 are 21.969, 36.234, 33.244, and 24.348 nm, respectively. The RMS values of GGr500, GGr600, GGr700, and GGr800 are 17.491, 38.027, 20.404, and 22.120 nm, respectively. Vertical distance denotes the thickness of graphene, and it is determined for GGr500, GGr600, GGr700, and GGr800 to be 7.298, 1.500, 7.406, and 27.695 nm, respectively. As the temperature increases, the thickness of the graphene products increases as well. The layer numbers were calculated via Equation (2). The layer numbers of GGr500, GGr600, GGr700, and GGr800 are 21, 3, 21, and 81, respectively. According to AFM results, the best results are obtained at 600°C temperature.
4. Conclusions
In this study, graphene was synthesized from graphite via the molten salt method. The synthesized graphene products were characterized by using XRD, UV-vis spectroscopy, DLS, electrical conductivity measurements, XPS, Raman spectroscopy, TEM, SEM, and AFM.
From the XRD results, all synthesized graphene products showed a peak that is characteristic of graphene and most of the synthesized graphene samples include 3 layers in average. These results are supported with the layer numbers that are obtained from Raman results; the ID/IG results show that most of the graphene products are few-layered. The UV-vis spectrum of GGr500, GGr600, GGr700, and GGr800 shows a peak around at 275-285 nm revealing the graphene structure.
According to the DLS results, the particle sizes of the graphene samples increased with increasing temperature. The electrical conductivities of GGr samples are 7-10 times higher than the electrical conductivity of CG. The XPS spectrum of GGr600 and GGr800 contains four peaks at ~284, 286, 290, and 285 eV, which correspond to the C=C, C-O-C, O-C=O, and C-C bonds, respectively. From the XPS results, it can be concluded that the carbon content of GGr decreases with increasing temperature. The Raman spectra consist of D, G, and 2D bands located at 1345, 1576, and 2659 cm−1, respectively. The synthesized GGr700 and GGr800 samples have a lower D band indicating that the ordered graphene structure included fewer defects and had good quality. According to ID/IG values, GGr600, GGr700, and GGr800 have a single layer, GGr500 has few layers, and RGO has multilayers. Graphene morphology was confirmed via both SEM and TEM studies. The average size of flakes observed in the SEM images was calculated as 2 μm. The TEM images show graphene structures with different thicknesses, ranging from 5 nm to 9 nm, indicating the presence of few-layer graphene. The layer numbers of GGr500, GGr600, GGr700, and GGr800 were calculated via AFM results as 21, 3, 21, and 81, respectively. Thus, it can be concluded that lower temperature gives better results.
Data Availability
The data used to support the findings of this study are included within the article.
Conflicts of Interest
The authors declare that there is no conflict of interest regarding the publication of this paper.
Acknowledgments
This work was supported by Gebze Technical University (GTU). The authors thank the Materials Science and Engineering Department of GTU for providing XRD measurements and particle size analyses and Advanced Technologies R&D Center of Bogazici University (BOUN) for XPS and Raman measurements. We would like to thank Editage (https://www.editage.com) for English language editing. Also, very special thanks are due to Bermak Kalip for lathe operations.