You are on page 1of 4

Introduction

The operation of a GaAs MESFET is similar to Si JFET, but the only difference is, in GaAs MESFET a metal semiconductor rectifying contact is used at the gate instead of a pn junction of a JFET [Sze-1981, Shur-1990]. A MESFET device is biased by applying two voltages, Vgs and Vds [Mass-1988]. These voltages are used to control Ids, which is present between the drain and the source of the device, by varying the electric field inside the channel. The field changes by changing the applied potential, giving rise to three distinct regions in the I-V characteristics of the device namely: (a) Linear Region; (b) Saturation Region and (c) Pinch-off Region. To simulate the dependence of Ids on Vgs and Vds, the field distribution inside the channel, which is dependent on device geometry, should be known [Rodriguez1994, McCamant-1990 and Ahmed (a)-1997]. Microwave MESFETs are of submicron gate length; called short channel devices having LG < a and usually fabricated by employing the gate recess technology [Das-1987, Bernstein-1998, Golio-1991 and Ahmed-1995]. Whereas low frequency devices having LG > a are called long channel devices [Adams-1993, Jaeckel-1986, Watts-1989]. This chapter, primarily, deals with mathematical description of I-V characteristics of both long channel and short channel MESFETs. It also describes the origin of intrinsic and extrinsic device parameters, which influence the performance of the device and are also required for its accurate modeling. 2.2. Long Channel Model

MESFETs fabricated by employing the condition LG > a are called long channel devices and the model that describes the behavior of such devices is termed as long channel model. It was first presented by Shockley in 1951 [Shockley-1951]. To describe the basic mechanism involved in the Shockley model, consider a crosssectional view of a GaAs MESFET shown in Figure (2.1). In this figure LG is the channel length, h is the depletion width, V(x) is the voltage drop underneath the gate, a is the epi-layer thickness, Wd and Ws represent width of the depletion at drain and source sides of the device respectively and A(x) is the available channel height defined as A ( x) = a h ( x) . (2.2.1) Here A(x) is a function of x because the electrons are flowing from source to drain

Figure (2.1): A cross-sectional view of a biased GaAs MESFET channel. b 2 s giving rise to a voltage drop V(x) along the channel which varies from zero at the source end to Vds at the drain side of the device. The potential difference between the gate and the channel at any point x is, therefore, given as V1 = b Vgs + V ( x) (2.2.2)

here b is the Schottky barrier height.

In order to calculate the width of the depletion layer formed by the Schottky barrier, it is assumed that the junction is abrupt and all the donor atoms (Nd) are ionized [Zeghbroeck-2004]. Under such conditions the two-dimensional Poissons equation can be written as [Zeghbroeck-2004] 2 x 2 2 + b y 2 = qN d s (2.2.3)

If we also assume that the variation of the potential along the length of the channel is negligibly small compared to y-directed field then Equation (2.2.3) can be modified as

b dE y = = qN d (2.2.4) y 2 dy

By solving Equation (2.2.4) one can evaluate the potential associated with the depletion region under the device boundary conditions as [Sze-1981]

q N h2 ( x) = d b 2 (2.2.5) s

When Vgs and Vds both are non zero then d is not uniform and it is function of x and can be written by using Equation (2.2.2) and (2.2.5)