The Field Effect Transistor (FET), which is the most basic building block of a majority of the modern ICs, has undergone many phases of development since its invention. By the time the FET was invented Silicon had already replaced Germanium as a result of the excellent properties of SiO2. The Metal Oxide Semiconductor (MOS) technology saw a transition from PMOS to NMOS to CMOS as a result of developments in fabrication. These developments were aimed at achieving mobility enhancements (transition from PMOS to NMOS) and reduction in power dissipation (transition to CMOS in digital circuits).
Graphene is a 2-D allotrope of carbon with atoms arranged in a honeycomb lattice, where carbon atoms are represented by red balls, as shown on the right. The existence of such a purely 2-D material before its isolation was highly debatable. In addition to electronics researchers, graphene has sparked the interest of physicists and materials scientists due to its other interesting properties.
The structure of a back-gated graphene FET is shown on the left. The graphene sheet serves as the FET channel and sits on top of a stack of dielectric and conducting material. The graphene sheet, the dielectric and conducting material form a parallel plate capacitor with the conducting material acting as the gate (G). The graphene sheet is connected to conducting contacts at the edges which serve as a Source-Drain (S-D) pair. The advantage of having a back-gated device is that the top surface of graphene makes conduction modulation via gas exposure possible. As a result, it can be used as a gas sensor.The goal of this project is to research deeper into the problem of mobility degradation due to water vapour adsorbed at the graphene-substrate interface. From mobility measurements from graphene on parylene [1], it is not possible to achieve theoretically predicted performances for graphene FETs on SiO2 substrates. Also, looking at the mobility improvement due to the transition from a hydrophilic to a hydrophobic substrate raises a natural question as to how far the performance of these FETs can be pushed by simply changing substrate properties. This project is an attempt to explore one of those properties, which is hydrophobicity, by experimentally determining the performance of graphene FETs on different hydrophobic substrates.
Polydimethylsiloxane (PDMS) was the first polymer investigated as it appeared to be a suitable candidate based on hydrophobicity. The surface free energy of PDMS is 19.8 dynes/cm which is lowest among polymers[2]. Also, for green light, the refractive index is in the range 1.44-1.46 which is very close to SiO2, at 1.47[3]. One other advantage of trying PDMS is that the recipes for making it are widely available in the literature. It is extensively used in research in the area of Microfluidics for microcontact printing and making microfluidic channels using PDMS moulds.
The recipe used in the experiments was borrowed from a research group[4]. PDMS was prepared by mixing the monomer solution with the curing agent from the Sylgard® 184 Silicone Elastomer Kit made by the Dow Corning company. The monomer and curing agent were mixed in a 10:1 ratio by volume. This mixture was then diluted 1% by weight in Heptane. The diluted solution was spun at 4500 RPM for 1 minute and allowed to cure. Two different curing methods were used for the same spin speed. One sample was cured by baking it at 100oC for 45 minutes and the other was allowed to cure in air (at room temperature) for 48 hours. Experiments were also performed by spin coating the wafers at different spin speeds and using different concentrations of the PDMS solution.
Experimental results have shown that it is possible to make films of PDMS that have thickness in the tens of nanometres. The first sample that was prepared by spinning at 4500 RPM for 1 minute was measured to have a thickness of 42 nm. These measurements were made using an ellipsometer. This result was very satisfying since it showed that it is possible to produce films of PDMS in the submicron range. The two samples that were cured differently showed almost the same thickness. Also, the thickness before and after curing was measured to be almost the same. The samples that were spun at different speeds (keeping all other steps the same) also showed approximately the same thickness. This result indicated that the spin coating method might not have an effect on the thickness. It was suspected that the reason for this was that the solution was too dilute.
By changing the concentration of the PDMS solution from 1% to 12%, a change in thickness was observed. In addition to that, a change in the thickness for different spin speeds was also observed for the 12% solution. But the exact relation between spin speed and thickness for the new concentration remains to be determined
REFERENCES
[1] S. S. Sabri et al., "Graphene field effect transistors with parylene gate dielectric," Appl. Phys. Lett., vol. 95, no. 24, Dec. 2009.
[2] (2010, June). http://www.surface-tension.de/ solid-surface-energy.htm
[3] Florian Schneider et al., "Process and material properties of polydimethylsiloxane (PDMS) for Optical MEMS," Sensors and Actuators A: Physical, vol. 151.
[4] Renée A. Lawton et al., "Air plasma treatment of submicron thick PDMS polymer films: effect of oxidation time and storage conditions," Colloids and Surfaces A, vol. 243, 2005
