Analysis of Polymer Field Effect Transistor
- PDF / 97,002 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 18 Downloads / 200 Views
Analysis of Polymer Field Effect Transistor Y. Roichman and N. Tessler1 EE Department, Technion Israel Institute of Technology, Haifa 32000, Israel 1
E-mail: [email protected]. Web page: http://tiger.technion.ac.il/~nir/
ABSTRACT We compare two basic organic FET structures both experimentally and theoretically. By using time resolved analysis we gain insight into the mechanisms affecting the performance of these structures. Using a two dimensional numerical model we focus on the top contact structure and demonstrate the difference between the two structures. INTRODUCTION Organic field effect transistors have been gaining attention over the past years. It seems that these transistors are out-breaking their performance and becoming very attractive for a range of applications. These are oscillators [1], flexible devices [2] smart cards/tags [3] small or large-scale [4], or even integrated optoelectronic devices [5, 6] [7]. Various device structures are often used to extract material properties with the common assumption that the unique structure of organic transistors behaves just like standard Si MOS-FET. Organic FETs have been simulated before [8, 9] however the emphasis there was on the band structure and transport properties of the material and not on the effects we are concerned with here. This paper describes numerical and experimental comparison of two relevant FET structures (Figure 1) and demonstrate that the top contact (TOC) structure is rather different. EXPERIMENTAL DETAILS Bottom Contacts Polymer Gold SiO2
Top Contacts Gold Polymer SiO2
Si
Si
Figure 1. Schematic description of top and bottom contact FET structures. Figure 1 describes the FET structures that are being considered. The active layer thickness is below 100nm. All I-V curves as well as time dependent curves were measured using HP semiconductor parameter analyser. In the numerical simulation the film thickness is 50nm and for the convenience of the simulation the insulator thickness was made the same. The mobility was assumed to be 5x10-3cm2V-1s-1. The drain-source distance (channel length) is 5 microns which is much larger then the film thickness. We have chosen a relatively high mobility and short channel length to reduce the computation time of our algorithm however, non-of the results presented here depend on these exact values. The size of each source/drain contact is assumed to be 1µm, again much larger then the film thickness. The simulation
C7.4.1
extends a few microns beyond the source drain contacts to reduce edge effects. The contact is modelled as thermionic emission controlled [10] with an injection barrier (φ) of 0.2eV. The boundary between the active layer and the insulator is taken as blocking for charges. When solving the Poisson equation we set the voltage (potential) at the gate source and drain electrodes and require that at all other boundaries of the simulation the electric field perpendicular to the boundary is set to zero. The simulation itself is a conventional 2Dsimulation that solves self-consistently both
Data Loading...
