Deposition of vertical carbon nanosheets by MPECVD at atmospheric pressure

The deposition process is studied of vertical carbon nanosheets by microwave (MW) plasma-enhanced chemical vapor deposition (PECVD) at atmospheric pressure. A coaxial MW plasma source with a surface-wave discharge at 2.45 GHz produces plasma in a gas mixture (Ar/H2/CH4) in the processing chamber. The emissive spectra of the plasma column in pulsed regime of the source are registered by an iHR550 spectrometer. The light from the plasma column is collected by a lens system connected to the spectrometer by an optical fiber. The dependence of the gas temperature in the plasma column on the absorbed MW power is obtained from the recorded OH-band and CN-band spectra by using the LIFBASE program. The plasma density is estimated from the Stark broadening of the Hβ-line, while the electron temperature is estimated by the line-ratio method using argon lines. The composition of the outlet gases from the chamber is measured and analyzed by an Agilent Micro gas chromatograph. The results obtained demonstrate the high efficiency of the methane decomposition process in the hot region of the plasma column (T g ~ 3000 K). The deposition of vertical carbon nanosheets is carried out at fixed plasma parameters and controlled substrate temperature. Cu plate and Ni-foam are used as substrates. The carbon nanostructures formed on the metallic substrates are studied by SEM; the dependence is thus obtained of their morphology on the plasma parameters, gas temperature and substrate temperature. The structures are confirmed as graphene sheets via Raman spectroscopy. The results demonstrate the viability of our system for deposition of vertical carbon sheets at atmospheric pressure.


Introduction
The carbon nanostructures (CNS) possess unique electrical, mechanical and optical properties making them an excellent choice for various applications. Depending on the number of nanoscale dimensions, the CNS are classified as 0D (nanoparticles), 1D (nanowires, nanotubes), 2D (nanowalls, nanosheets) or 3D structures. Carbon nanowalls and vertically standing few-layer graphene sheets are of great interest due to their specific properties and morphology (intersecting vertical sheets with sharp edges and high specific surface area).
Carbon nanowalls (CNW) have potential applications in microelectronics as sensors, field emitters and electrodes for supercapacitors. Promising methods for graphene production are the plasma-enhanced chemical vapor deposition (PECVD) [1][2][3] and the free standing method [4][5][6][7]. A variety of gas-discharge plasmas (DC, RF or MW) can be used for this purpose. Synthesizing CNS at atmospheric pressure in Ar/CH4 or N2/H2/CxHy gas mixtures allows for shorter deposition times and easier 1 To whom any correspondence should be addressed. technological implementation compared to the typically used low-pressure CVD methods. The MW discharges are usually characterized by a high-density plasma and relatively high gas temperature, which provides efficient decomposition of hydrocarbons and production of CxHy radicals and C2.
In this study, an MW plasma reactor is used for CNS deposition at atmospheric gas pressure under controlled conditions. The plasma parameters are determined and the structures deposited are analyzed by SEM.

Experimental set-up and diagnostics
The experimental setup (figure 1) has been discussed in other publications [8,9] and has undergone a few modifications which will be explained below. A low-power MW discharge in a ceramic tube is used to create plasma in a gas mixture of Ar, CH4 as a precursor and additional H2 to improve the deposition process. An MPG-4M generator is used to feed a signal at 2.45 GHz to an MW plasma exciter through a coaxial line. The MW exciter includes alumina ceramic as a dielectric medium and also as a surface-wave discharge tube. The plasma source and the substrate holder are placed in a quartz vessel with metal flanges filled with a gas mixture of Ar, H2 and CH4. The discharge is ignited by the MW exciter and the substrate (nickel foam, Cu-plate) is heated by both the plasma jet and an additional heater. Monitoring and control of the substrate temperature are performed via a thermocouple electronic system. The reflected and forward power is measured by an HP 437B Power Meter through a Pasternack PE 2219-30 directional coupler; the data received from it is used to achieve a better matching of the plasma source by a Maury 1878C triple stub. The pulse parameters of the MW signal are measured by a Tektronix TDS 360 oscilloscope.
The setup has two main improvements over the one used previously [8,9]. First, an Agilent Technologies 490 Micro gas chromatograph is used to measure the outlet gases from the process chamber. Second, the emissive spectra of the plasma flame are measured by both an HR Ocean Optics spectrometer and a new Horiba iHR550 imaging spectrometer. The light from the plasma flame is collected by a collimator and fed to the spectrometers through an optical cable. The data obtained is used to determine the rotational gas temperature using the LIFBASE program to fit and compare simulated OH and CN-bands to the experimental ones [10]. The electron temperature and electron density are obtained using the line-ratio method [12] and the Hβ line broadening [13]. Also, the presence of certain bands in the plasma spectrum is used to determine that the successful production is taking place of reactive species usually associated with the deposition process.

Results
Using the LIFBASE software, we fitted the simulated to the experimental spectra to determine the gas temperature (Tg) of the plasma [10]. The parts of spectrum used were the OH (A -X) 0-0 band (305 nm -315 nm) and the CN (B2 Σ+ -X2 Σ+ ) 0-0 band (387.2 nm -388.4 nm) [11]. Due to the necessity for controlled environment and pre-treatment of the substrate, the processing chamber was initially filled with Ar/H2 gas mixtures. The oxygen amount in the chamber strongly decreased as the pre-treatment time and the heater temperature were increased; the OH emission band intensity decreased accordingly. After CH4 was added to the gas mixture and the methane was pyrolyzed in the plasma column, the CN band intensity increased; this band was used to determine the gas temperature.
In the deposition process of vertical nanosheets, a gas mixture Ar/H2/CH4 with volume ratios 180/4/1 was used. In this process, the argon is a working gas for the plasma source, H2 serves as a catalyst and CH4 is decomposed in the plasma column and is the source of hydrocarbon, C and C2 species.
The spectra were used to measure the rotational gas temperature at different substrate/plasma exciter distances (5 -10 mm) and at different input MW power (6 -14 W). The results show that as the power is increased and the tube-substrate distance is decreased, the gas temperature rises from 2800 K up to 3800 K. The presence of the intensive swan band in the measured spectra at gas temperatures above 3000 K confirms that the methane decomposition and C2 radicals production are efficient ( figure 2).
The argon lines from the spectra were used to determine the electron temperature (Te) and electron density (ne) via the line ratio method [12] and the Hβ line broadening [13] as a second method to obtain ne. During the deposition of graphene nanostructures under the conditions quoted, the plasma parameters were determined to be ne ≈ 4.5 -6×10 20 m -3 and Te ≈ 1.1 eV.
The graphene nature of the deposited nanostructures as a thin layer on the nickel-foam wires was confirmed by Raman spectroscopy ( figure 3) showing the 2D, D+D' and 2D' peaks.
SEM imaging was used to study the nanostructures' morphology. It shows that the structures are carbon nanowires when a low power (6 -9 W) and a large substrate/plasma exciter distance (10 mm) are used. Increasing the power (10 -14 W) and decreasing the distance changes the morphology of the structures into vertical carbon nanosheets (figure 4). Decreasing the distance even further (< 5 mm) leads to stronger plasma etching processes and removes most depositions from the substrate. The average deposition time for these structures is 8 min; exceeding this time results in the deposition of a thick carbon layer with no visible nanostructures.
Both types of structures required a sufficiently high substrate temperature in the deposition process, which was maintained by a cartridge heater; controlling the temperature within the 600 -700 C resulted in a successful deposition.

Conclusions
The results are presented of PECVD using an MW plasma source on nickel foam substrate. They demonstrate the validity of the method for controlling the deposition process allowing production of different graphene nanostructures. The key advantage of the described method is the possibility to deposit nanosheets and nanowires at atmospheric pressure. The power requirements of up to 14 W guarantee the cost-effectiveness of the devised system.