ABSTRACT
We achieved a small resist thickness as thick as 100 nm, which will define the Γ-gate foot, by using a mixing liquid of PMMA and MCB (Monochlorobenzene) (1.5 : 1) and a coating speed of 7000 rpm. A variation of PMMA thickness according to the mixed ratio of PMMA and MCB is given in Fig. 2. In order to obtain a reproducible 0.1 μπι Γ-gate, we performed an experiment to observe the variation of the gate length caused by the change of the development time and the exposure dose. The exposed wafer was developed in MIBK : IPA (1 : 3). We determined optimum fabrication conditions for a development time of 140 seconds and a dose of 900 μΟ/αη2 through these experiments. Fig.3 represents a variation of line width as a function of electron beam dosage and development time and Fig. 4 shows the SEM micrograph of a cross-sectional view of PMMA with variation in line profiles as a function of electron dosage at a constant development time of 140 seconds.. The top two layers PMMA/P(MMA-MAA) that have different sensitivity will form a reliable overhang structure for metal liftoff. The linewidth of the top two layers determines the cross-sectional width of the cap. Using the fact that the cross section of the cap can easily be increased by overdevelopment, we experimentally found an appropriate dose of 90 μΟ/αη2 as a result of conducting the exposure dose test ranging from 60 μηι/cm2 to 105 μηι/cm2 in 15 μ(ϋ/αη2 increments. After being exposed to a dose of 90 μ(2/αη2, the top layer (PMMA) was developed using MCB for 15 seconds, and then the middle layer of P(MMA-MAA) was developed using the mixed solutions of methanol and IPA (1 : 1) for 28 seconds. The degree of undercut can be controlled by the development time. Fig. 5 shows a SEM photograph of the gate head fabricated by using this procedure. Selective etching is an important step in processing GaAs/AlGaAs structures because the etched depth is limited within a few hundred angstroms and critically affects the threshold voltage of devices. In this report, the citric acid solution/H2C>2 and citric acid solution /H2O2/H2O are compared. The samples of undoped
Fig. 6 AFM and SEM images taken from the surfaces etched by a solution of (a)50 % citric acid solution/H20 2 (1:1), (b) 1:2, (c) 1:3, and (d) 50 % citric acid solution /H20 2/H20 (1:3:1).
