Enhancing Growth of Multi-Layer Graphene Synthesis on Glass Substrate Though Ni Catalyst Annealing Using Hot Wire in Plasma Very High-Frequency PECVD Method

This study synthesized multi-layer graphene on a glass substrate using the high-frequency plasma-enhanced vapor deposition method with a hot wire in plasma. In the growth process, plasma was generated from methane gas (CH4) using an RF power generator at 70 MHz frequency and nickel (Ni) as the catalyst, which is deposited and annealed beforehand. This study aims as a preliminary to obtain optimum parameters for MLG growth using the HW IP-VHF-PECVD method. Raman spectroscopy, utilizing a 532 nm laser and an 1800 g/mm grating, detected distinctive D-band, G-band, and 2D-band peaks at wave numbers of 1335.31, 1607.74, and 2660.99 cm−1, respectively, in unannealed catalyst samples. Raman analysis, incorporating the I2D/IG ratio, revealed the presence of multi-layer graphene exhibiting diverse ratios (0.07 – 0.22) and crystal sizes (15.62 – 20.70 nm). SEM analysis demonstrated enhanced homogeneity in grain size and uniformity in thickness following Ni catalyst annealing. EDS confirmed the successful growth of graphene with carbon identified as the primary element. The annealing process at 400 °C for two hours resulted in graphene with a higher mass percentage and a more significant percentage of carbon atoms. These findings underscore the potential of the HW IP-VHF-PECVD method for producing multi-layer graphene, particularly in the context of solar cell applications, with further optimization of parameters.


Introduction
In the 21st century, graphene has become a rapidly growing primadonna in materials science because graphene has a high surface area (2,630 m 2 g −1 ), high intrinsic mobility (200,000 cm 2 v −1 s −1 ), large Young's modulus (~1.0 TPa), thermal conductivity of (~5,000 Wm −1 K −1 ), as well as good electrical conductivity (2.1 x 10 -4 S cm -1 ) [1].Graphene itself is a material with carbon atoms bonded together to form a 2D (two-dimensional) sheet with a hexagonal structure resembling a honeycomb [2], consisting of two sublattices of carbon atoms, each of which has a σ (sigma) bond [3].In 3D (three-dimensional) form, the arrangement of graphene is called graphite.Graphene is categorized into three types based on its number of layers, namely single layer, double layer (bilayer), and several layers (few layers), commonly referred to as multi-layer graphene (MLG), with the condition that the number of layers (n) d 10 [1,4].Graphene with more than 10 layers is precisely termed graphite.
So far, the chemical vapor deposition (CVD) method is the most widespread and promising method for growing graphene, including MLG with a large surface area.Many investigations have been carried out to analyze the application of the chemical vapor deposition (CVD) technique in the growth of graphene for a wide range of uses, with particular attention given to its efficiency in the domain of solar cells [5][6][7][8][9][10].These studies haven't only examined the ordinary CVD method but explored various changes, including plasma-enhanced chemical vapor deposition (PECVD) [11,12].Regarding the use of this method, the electronic materials physics expertise group (Fismatel) of the Bandung Institute of Technology, PECVD laboratory unit, has so far succeeded in growing several layers of graphene (FLG) on a glass substrate with silver (Ag) metal precursors [13], using the hot wire cell-very high frequencyplasma enhanced chemical vapor deposition (HWC-VHF-PECVD) method.The results show that MLG has been successfully formed based on scanning electron microscopy (SEM) images with an area of approximately 0.5 m 2 deposited on previously grown silver catalyst grains.However, the quality of the graphene produced is still relatively low and requires further optimization of growth parameters.In this investigation, Ni is chosen as a catalyst rather than other metals like Ag and Cu based on several factors contributing to its superior performance.One crucial aspect is its remarkable carbon solubility, facilitating further carbon deposition after cooling [14].Moreover, Ni exhibits a higher reactivity due to its enhanced lattice compatibility with MLG, as established by previous studies [15,16].Moreover, the selection of Nickel as the catalyst is further supported by the research findings of Lahiri et al. [17], who examined the interaction between MLG and Nickel (111) and observed that narrower cluster width values were associated with increased adhesion, as evaluated from electron microscope images.In comparison, broader cluster width values led to decreased adhesion.Moreover, utilizing Ni as the catalyst is unlikely to result in the rapid formation of high-quality graphene directly on glass substrates at a relatively low temperature of 700 ºC, thus not resulting in exceptional transparency and favorable electrical conductivity.
Based on MLG's achievements and promising potential in increasing the conversion efficiency of thin film solar cells [12,18,19], in this study, the growth of MLG was carried out on a nickel (Ni) metal catalyst, which will later be applied to solar cell devices based on polycrystalline silicon (p -Si) with pi-n type.This study aims as a preliminary to obtain optimum parameters for MLG growth using the HW IP-VHF-PECVD method.It is fundamental because the characteristics of MLG can be used as a window layer and back reflector layer in thin-layer solar cells based on p-i-n type silicon, especially p-Si, which has not been widely reported.In this study, the MLG layer was grown using the hot wire in the plasmavery high-frequency PECVD (HW IP-VHF-PECVD) method, which is a modification or development of the conventional PECVD method.The modification made was by changing the RF power generation frequency from 13.56 MHz to 70 MHz and adding a hot filament in the form of a tungsten wire at the gas system input parallel to the substrate so that the gas used as a precursor would be better and more radical.The hot filament decomposes the gas molecules before entering the electrode area.

Experimental methods
Before growing the MLG layer on the Ni catalyst, both annealed and unannealed Ni catalyst samples were cleaned from dust that might stick to them using a nitrogen gas gun.After that, the sample was then inserted into the PECVD reactor chamber.In this research, the MLG layer was grown on an annealed Ni catalyst for two and four hours at temperatures of 350, 400, and 450 o C, respectively, with a hydrocarbon precursor in the gas phase, namely methane gas (CH4), where, the precursor gas (CH4) which enters the chamber by first hitting a hot filament (hot wire), then by utilizing this hot filament, the CH4 gas will be decomposed into a number of its constituent components in the form of molecules, ions or more superficial elements, so that plasma is formed.A chemical reaction occurs which separates the C and H atoms.The grains on the catalyst substrate then guide the C atoms to form MLG. So, apart from the decomposition process resulting from applying high frequencies, decomposition is also caused by hot filaments.It will lower the growth temperature and make the decomposition process more efficient.The plasma resulting from this decomposition will then deposit carbon atoms on the catalyst's surface to form an MLG layer.Growth parameters of the MLG layer, such as source gas type, gas flow rate, chamber pressure, power, and deposition time (calculated from the moment the substrate cover shutter is opened), are presented in Table 1.
To identify the type of peak of the MLG layer grown on the sample substrate of the previously annealed Ni catalyst layer, we characterized the sample using a Raman spectrometer (HORIBA -The LabRAM HR Evolution Raman Microscopes) with a 532 nm laser beam and 1800g/mm grating in the materials characterization laboratory at the Integrated Laboratory and Research Centre (ILRC), University of Indonesia.Additionally, the morphological structure and composition of the MLG layer were analyzed using SEM-EDS (JEOL JSM-6510 LA instrument at the SEM Laboratory, FMIPA Integrated Laboratory, ITB) at 30,000x magnification.

Raman spectroscopic characterization
The results of Raman spectroscopic characterization using laser light with λ of 532 nm (2.33 eV) in Figure 1 show that a graphene layer has been successfully grown on the surface of the Ni catalyst produced in this study.For comparison, the results of the significant Raman spectra feature of graphene and graphite produced in other studies [20,21] are also shown in Figure 2. By comparing the Raman spectra features of graphene samples grown in this study, it was found that they match the Raman spectra of graphite in Figure 2, namely 1347, 1596, and 2647 cm -1 , which are associated with D-band, G-band, and 2D-band, respectively [22].
The resulting graphene layer, as in Figure 1, shows the presence of typical peaks of graphene, namely D-band, G-band, and 2D-band (in other studies, also called G'-band) each at wave number 1335.31, 1607.74, and 2660.99 cm -1 , respectively for graphene grown on a non-annealing Ni catalyst layer.Other peaks, such as the D*-band and 2D"-band, respectively, were found at 2911.71 and 3205.26cm -1 wave numbers.When compared with the results of other studies, the peak positions of the G-band and 2Dband graphene that were successfully grown in this study experienced a shift when compared with the positions of the G-band and 2D-band of the graphene layer, which was also grown using a glass substrate, as obtained Wang et al. [23], namely 1582.5 and 2672 cm -1 respectively.In graphene, the shift in Stokes phonon energy caused by laser excitation leads to the occurrence of two prominent peaks in the Raman spectrum, namely G, which represents a primary field of vibrational modes, and 2D, as well as an additional second order of different vibrations, D. For lasers with λ = 532 nm gives the Raman spectra G, 2D, and D peaks at wave numbers 1580, 2690, and 1350 cm -1 respectively [24].The D and 2D positions are dispersive (depending on the laser excitation energy) [20], so using lasers with different λ will produce different spectral peaks.Apart from that, the shift in the peak of the Raman spectra can be caused by several factors, such as defects caused by residual particle contamination remaining in the growth chamber, abundant edges, and boundary effects formed during the deposition process [22,25].
Because this research aims to produce MLG, the number of graphene layers successfully grown in this experiment was also determined.Determination of the number of graphene layers formed in this study, apart from referring to the data presented in Figure 2 c, is also can be seen from the ratio of the   2D peak intensity (I2D) to the G peak intensity (IG).The I2D/IG ratio can provide information about the number of graphene layers formed, as it is known that in graphene, there are 2D-band Raman peaks associated with complex electron interactions.This peak is formed due to Raman scattering involving two photons.The intensity of the 2D-band peak depends on the number of graphene layers and the graphene quality.The G-band Raman peak is related to the sp 2 vibration of carbon that is neatly arranged in the graphene structure.The G-band peak intensity can be used as a reference to compare with the 2Dband peak [26,27].By comparing the 2D-band peak intensity with the G-band peak intensity (I2D/IG), an indication of the number of graphene layers formed will be obtained.A higher I2D/IG ratio tends to indicate more layers.
In graphene research, the ideal I2D/IG ratio is usually around 2, which indicates the formation of an excellent monolayer graphene structure.A general estimate for the number of graphene layers is based on the I2D/IG ratio of the Raman spectrum [28].A general estimate for graphene layers based on the I2D/IG ratio consists of (1).Monolayer (one layer), where the I2D/IG ratio in monolayer graphene tends to have a higher value, often more than two or even higher.It is because the 2D peak in monolayer graphene has a significant intensity compared to the G peak, (2.) bilayer (two layers), the I2D/IG ratio in bilayer graphene is usually lower compared to monolayer graphene but is still relatively high, in the range of 1-2, and (3) multi-layer (more than two layers), where the number of graphene layers increases, the I2D/IG ratio tends to decrease.In graphene with more than 2 layers, the I2D/IG ratio can range from 0.8 to 1.5, depending on the number of layers and the degree of defects in the structure [22,24,25].Peak positions and calculation of ID/IG and I2D/I2G ratios are shown in Table 2.
Furthermore, the ID/IG ratio can also be used to determine the size of graphene crystals using the Tuinstra-Koenig equation [16,29,30].
where La represents the size of the graphene crystal, λl is the laser wavelength used in Raman spectroscopy in the experiment (532 nm).With the help of the Raman Spectroscopy Crystallite Size Calculator [31], the graphene crystal size was obtained, as presented in Table 2.

Scanning electron microscope (SEM) analysis
SEM characterization results of the graphene layer grown on an unannealed Ni catalyst show an uneven surface morphology.As shown in Figure 3 (a), there are empty parts with a distribution of grains shaped like corn kernels.It is confirmed through the cross-sectional image; some parts are in the form of depressions, and some peaks resemble hills, which indicate the thickness of the layer.Using the ImageJ    [32], an average graphene grain diameter size of 17.50 nm was obtained with an average layer thickness of 44.50 nm.
Different things were obtained for the graphene layer grown on the Ni catalyst layer, which was annealed at temperatures of 350, 400, and 450 o C within two and four hours.As presented in Figure 3 (b) -(d), it can be seen that the surface morphology of graphene grown on a Nickel catalyst layer that was annealed for two hours has a smoother surface shape than graphene grown on a non-annealed Ni catalyst layer, with more evenly distributed grains.The size appears more extensive and more uniform/homogeneous.The average diameter of these grain sizes is 24.50, 33.50, and 22.00 nm, respectively, for graphene grown on annealed Ni catalyst layers at temperatures of 350, 400, and 450 o C. Apart from the larger grain size, from measuring the thickness of the SEM, A larger size was also obtained in a cross-sectional image with thicknesses of 67, 69, and 54.50 nm, respectively.
The same results are also shown in the SEM image of the surface of the graphene layer, whose catalyst layer was annealed for four hours, as shown in Figure 3 (d) -(f).However, at an annealing temperature of the catalyst layer of 450 o C, the number of grains scattered on the surface of the resulting image appears to be much less.Also, their size is smaller than graphene, whose Ni catalyst layer is not annealed.Figure 3 (g) shows that the grain size is more significant and looks brighter, but it is suspected that this is a charging effect during the characterization process.This condition was confirmed by EDS results, where in this sample, the number of atoms and Ni %were much less than in other samples, including samples that were not annealed.With a smaller mass % and atoms % due to the annealing process in the catalyst layer, resulting in fewer C atoms from the precursor gas used (CH4) interacting and bonding with Ni atoms during the graphene growth process, reducing the number of bonds between C atoms.formed during the growth process and the grain size distribution of the graphene grown is also smaller, which ultimately causes the layer formed to be thinner than the other samples.However, this relationship is general and can vary depending on the growing method and experimental conditions during the growing process.For this reason, in the subsequent research stage, the graphene growth parameters will be optimized so that the graphene obtained is suitable for its application for p-i-n type polycrystalline silicon solar cell devices.The overall calculation results of the average grain diameter size and thickness of the graphene layer produced in this research are presented in Table 3.
From the results of this SEM characterization, it can be seen that the optimum conditions for the Ni catalyst layer produce graphene with a uniform/homogeneous size and relatively consistent uniform grain size, as well as a layer that is neither too thick nor too thin These optimal conditions are observed in samples annealed for two hours at all three temperatures.Additionally, when the graphene grains are smaller, the thickness of the graphene will tend to be more consistent and uniform.This means the graphene thickness in the observed SEM image area will be more homogeneous and less variable.

Energy dispersive X-ray spectroscopy (EDS) analysis
Apart from being shown through the Raman spectra results, the success of graphene growth in this experiment was also shown by the results of the EDS characterization of the samples.As shown in Table 4, the non-annealed Ni layer sample contains trace C with a mass % of 6.63, which is the main element of graphene.Apart from that, other elements were also found, such as Ni with a mass % of 16.81, which came from the catalyst layer used in growing graphene, as well as the elements O, Si, Na, and Ca, which were elements from the glass substrate used with a mass % of 36.44.27.31, 6.12, and 4.07, respectively, as well as Mg and Al elements, which are thought to originate from contamination during the growth of the Ni catalyst layer.The same results were obtained for graphene grown on a Ni catalyst layer that had been annealed for two and four hours at temperatures of 350, 400, and 450 o C, respectively, as presented in Table 4.This data shows that the annealing process on the Ni catalyst layer will increase the number of carbon atoms deposited on the catalyst's surface.Compared with graphene, whose Ni catalyst layer is not annealed, the mass % and atomic % of C atoms in graphene whose Ni catalyst layer is annealed at temperatures of 350, 400, and 450 0 C for two and four hours, respectively, are much higher.Based on the data in Table 4, when compared to graphene whose Ni catalyst layer was not annealed, the most The same thing also happened to graphene that was annealed for four hours, where an increase in the mass % and atomic % of C atoms when compared to the graphene sample whose Ni catalyst layer was not annealed occurred in the sample whose Ni catalyst layer annealed at temperatures of 350 and 400 0 C with a percentage increase of % mass respectively 79.94 and 90.65% and 79.09 and 83.06% for the increase in the number of atoms %.Meanwhile, for the graphene sample whose Ni catalyst layer was annealed at a temperature of 450 0 C, when compared with the graphene sample whose Ni catalyst layer was not annealed, the mass % and atomic % of C atoms decreased to 31.07 and 36.20%,respectively.The decrease in mass % and atomic % of C atoms in graphene samples grown on a Ni catalyst layer annealed for four hours at a temperature of 450 0 C is thought to be caused by a longer annealing time on the Ni catalyst layer resulting in degradation of the catalyst layer, thereby reducing the number of carbon atoms that can be absorbed or supports the growth of graphene.This also affects the kinetics of the reaction between the carbon precursor (CH4) and the Ni catalyst layer, thereby influencing the efficient amount of carbon atoms absorbed and incorporated into the catalyst layer.In addition, a longer time and higher temperature during the annealing process can result in the rate of atomic diffusion in the material layer being slower, resulting in a more significant reduction in the atomic element content of a material [33,34], which also affects precursor atomic interactions.CH4 with Ni as the catalyst layer [23].This can be seen from the results of measuring the grain size and thickness of the graphene layer in this sample, as presented in Figure 3 (g) and the data in Table 3.However, this assumption is still general and can vary depending on growth methods and experimental conditions during the growth process.Therefore, in the following research stage, graphene growth parameters will be optimized by controlling the precursor gas flow rate, PECVD reactor chamber pressure during growth, RF power, substrate temperature, and tension on the hot wire so that the graphene obtained is suitable for its intended purpose-application for the use of p-i-n type polycrystalline silicon solar cell devices.

Conclusion
Optimum conditions for graphene growth were obtained by annealing the Ni catalyst layer for two hours at a temperature of 400 °C.The graphene produced under these conditions has better morphology, larger and more uniform grains, and a high mass percentage of C. Raman spectroscopy results show the growth of graphene layers on the Ni catalyst, with D-band, G-band, and 2D-typical band, where the results of Raman spectroscopy analysis show that the graphene grown in this study is a multi-layer structure, with an I2D/IG ratio ranging from 0.07-0.22.Then, using the Tuinstra-Koening equation, the crystal size of the growing graphene layer is estimated to vary between 15.62 -20.70 nm.Furthermore, SEM analysis revealed that the annealing process on the Ni catalyst layer caused an increase in grain size homogeneity and thickness uniformity of the graphene layer grown on the Ni catalyst layer compared to graphene grown on the unannealed Ni catalyst layer.The non-annealed sample has an uneven morphology with non-uniform grain distribution, while graphene grown on a Ni catalyst layer that is annealed for two hours has a smoother morphology with larger and more uniform grains than the non-annealed sample.Furthermore, EDS analysis confirmed the success of graphene growth by detecting the element C as the main element in graphene, where increasing the annealing temperature of the Ni catalyst layer increased the number of carbon atoms deposited on graphene, with a higher mass percentage and percentage of C atoms in graphene annealed for two hours at 400 0 C.

Figure 1 .
Figure 1.The Raman spectra of graphene grown on a Ni catalyst layer.

Table 1 .
MLG layer growth parameters

Table 2 .
Raman spectra peak positions, ID/IG and I2D/IG ratios, and type and size of graphene crystals

Table 3 .
Grain size and thickness of the graphene layer

Table 4 .
Results of EDS characterization of the surface of the graphene layer