Selective area multilayer graphene synthesis using resistive nanoheater probe

Graphene has been a material of interest due to its versatile properties and wide variety of applications. However, production has been one of the most challenging aspects of graphene and multilayer graphene (MLG). Most synthesis techniques require elevated temperatures and additional steps to transfer graphene or MLG to a substrate, which compromises the integrity of the film. In this paper, metal-induced crystallization is explored to locally synthesize MLG directly on metal films, creating an MLG-metal composite and directly on insulating substrates with a moving resistive nanoheater probe at much lower temperature conditions (~ 250 °C). Raman spectroscopy shows that the resultant carbon structure has properties of MLG. The presented tip-based approach offers a much simpler MLG fabrication solution by eliminating the photolithographic and transfer steps of MLG.


Resistive Nanoheater Probe Calibration
The figure below (Fig. S.2) shows the arrangement used to calibrate the resistive nanoheater probe. The 12 µm thermocouple (CHAL0005, Omega Engineering, Norwalk, CT, USA) is mounted in a fixed position on top of a metallic hand with a magnetic base. The resistive nanoheater probe is mounted on a soft sponge placed on top of a 3-axis manual micromanipulator with a magnetic base. The micromanipulator allows movement of the probe in the x, y, and z directions. First, the probe and the thermocouple are placed close enough until they are visible through the camera. Then, the probe is connected with alligator clips to a source meter (Keithley 2400, Tektronix, Inc., Beaverton, OR, USA) to feed voltage to the resistive nanoheater probe. Once everything is ready, they are zoomed in until the thermocouple is clear on the screen. The probe is then moved through the micromanipulator in all the necessary directions to make contact with the thermocouple. With the source meter, the voltage is increased to raise the temperature of the tip, and its value is recorded, as shown in

Experimental Setup
The figure below shows the schematics of the homemade system for multilayer graphene in-situ synthesis, detailed in the methodology section of the paper.

Supplementary Figure S.3.
Homemade system schematic for in-situ synthesis detailing each component of the system. The system consists of a tailored glass chamber with a flat top to be used as an optical window. Each fitted connector of the chamber serves a purpose, such as pressure monitoring, bidirectional gas flow, vacuum, and electrical feed. A fixture and a movable stage comprise the interior of the chamber in order to hold the resistive nanoheater probe and to maneuver a sample in the XYZ direction, respectively. The LabView program in the computer allows the user to control the platforms for precise movement of the sample to touch the probe for synthesis.

Raman Spectra Evolution
The Raman spectra of the samples heated at different times are shown in Figure S Once the area of interest is locally heated for one minute, the separation between the D and G peaks becomes noticeable in the Raman spectra, as seen in Fig. S.4b. The G peak shifts towards a higher value, 1578 cm -1 , and the intensity of the D peak shifts to 1352.12 cm -1 . When a-C is heated after deposition, it follows an ordering trajectory from a-C to graphite following stage 2 in the FR model 3 . In the presence of thermal energy, the clustering of sp2 sites increases into ordered aromatic rings, and the average distance C=C shortens 3,4 . The intensity D-to-G ratio (ID/IG) obtained in this phase is 0.80. The appearance of the 2D peak indicates graphitic development for which the intensity G-to-2D ratio (IG/I2D) is 1.74. At this stage, the D and G peaks show to have FWHM of ≈335 cm -1 and ≈105 cm -1 , respectively. The 2D peak band measures an FWHM of ≈85 cm -1 .
After continuous heating of 15 minutes, the intensity of the D peak in the Raman spectrum rises at almost the same level as the intensity of the G peak, as seen in Fig. S.4d, consequently increasing the ID/IG ratio to 0.97. On the other hand, the IG/I2D ratio is calculated to be 1.50. However, the D peak does not shift much, remaining at 1333 cm -1 , whereas the G peak shifts upward to 1584 cm -1 , and the 2D peak shifts downward to 2662 cm -1 . Furthermore, the D' peak subtly appears ≈1620 cm -1 as a shoulder with the G peak that is hardly detected. The ID/IG ratio increases and the strengthening of the D peak indicates an increase in the number of defects 5,6 . The possible activation of the D' peak suggests an amorphization trajectory, namely disordering, in the Raman spectrum of the FR Stage Model 1 . Additionally, the G and 2D peaks slightly increase their width while the D peak is decreased, showing FWHM of 83 cm -1 for the D band, 85 cm -1 for the G band, and 87 cm -1 for the 2D band. Between the D and G peaks, there is an unidentified small peak that is irrelevant to the MLG structure characterization, and therefore its origin is not examined. The Raman spectra of virgin Sn show characteristic peaks below 700 cm -1 and no presence of peaks above this value 7 . However, it has been revealed that the Raman spectrum for SnO develops other peaks when annealed, but no study has been found showing peaks beyond 700 cm -1 8 . Therefore, the peak in between the D and G may come from the metal when annealed at 250 °C or a yet to be identified source.
When the sample is heated for 60 min, not only the D, G, and 2D peaks intensify, but they also widen, as shown in Fig. S.4e. First, the D peak maximum intensity surpasses the intensity of the G peak, consequently increasing the ID/IG ratio to 1.10. Additionally, IG/I2D ratio is found to be equal to 1.94. The D and the 2D peaks shift downward to 1327 cm -1 and 2644 cm -1 , respectively. Meanwhile, the G peak remains at 1581 cm -1 . The FWHM of the peaks are ≈95 cm -1 for the D band, ≈105 cm -1 for the G band, ≈118 cm -1 for the 2D band, confirming the increase of their width. The broadening of the G peak is attributed to the intensification of the D' peak, which merges with the G peak. The activation of the D' peak is due to defects causing intravalley double-resonance process 9 . For the sake of simplicity, usually, these two peaks are conveniently considered as one broad single G peak when the separation between them is not apparent 10 . Moreover, an additional visible peak appears ~ 2900 cm -1 with an FWHM of ≈200 cm -1 , which corresponds to the (D + D') peak 9,10 . The resultant Raman peaks at 60 minutes confirm that the spectra evolution follows the FR Model stage 1 disorder 1,3,10 . Therefore, the continuous heating of the region breaks the crystal symmetry, increasing the defect density and introducing disorder into the MLG. It is worth noting that disorder can also be induced by sp3-defects which usually originate from chemisorptions, also referred to as chemical defects, where there is a change of hybridization (from sp2 to sp3) such as oxidation [11][12][13] .