Scaling down of silicon-based transistors is facing various problems such as increase in the gate leakage and power consumption. As a solution to resolve those problems, III-V compound semiconductors are introduced as a channel material. Because III-V compound semiconductors have outstanding electron transport properties, they can enable the downscaling of transistors [1,2]. However, the introduction of III–V compound semiconductors into Si-based infrastructures and manufacturing processes still have many problems to be solved. One of the largest problems is the removal of photoresists on III–V compound semiconductors. In the general Si-based manufacturing process, sulfuric acid and hydrogen peroxide mixture (SPM) or ozonated water is used to remove photoresist. However, such chemicals may not be used on III–V compound semiconductors because they etch or oxidize III-V compounds [3]. Also, photoresist removal process should be achieved at low temperature to avoid the surface damage of semiconductor. Therefore, photoresist removal process using organic solvents at room temperature that cause no material loss of III-V compounds was conducted in our study.

Both ArF photoresist on GaAs wafer and KrF photoresist on trench patterned GaAs wafer were tested. Implantation was conducted with P ions on both photoresists at doses of 5 × 10¹³, 5 × 10¹⁴, and 5 × 10¹⁵ atoms/cm² at 30 keV and 10^-6Pa. Implanted ArF and KrF photoresists on GaAs wafer were removed using two different organic solvents in sequence to reduce photoresist removal process temperature. Low molar volume solvents (formamide, acetonitrile, nitromethane, and monoethanolamine (MEA)) were tested as the first step solvent in the two-step photoresist removal process and dimethyl sulfoxide (DMSO) which has high affinity to photoresist was selected as the second step solvent. The implanted photoresists on GaAs were dipped in each solvent for various times from 5 to 30 min. In order to evaluate the photoresist removal efficiency, field-emission scanning electron microscopy (FE-SEM) and optical microscope were used. Also, to analyze the reactivity of organic solvents with GaAs, X-ray photoelectron spectroscopy (XPS) and atomic force microscopy (AFM) were conducted.

In case of formamide + DMSO two-step treatment, while ArF photoresists implanted at doses of 5 × 10¹³ and 5 × 10¹⁴ atoms/cm² were removed, photoresist implanted at 5 × 10¹⁵ atoms/cm² could not be removed, even after dipping in each solvent for 30 min. Nitromethane and MEA + DMSO showed similar results with formamide + DMSO. However, two-step treatment with acetonitrile + DMSO resulted in complete removal of ArF photoresist implanted at doses of 5 × 10¹³_, 5 × 10¹⁴ and even 5 × 10¹⁵ atoms/cm² after dipping for 15 min in each solvent (Fig.1). The behavior of photoresist removal in different organic solvents can be explained by combining the permeability of solvents through photoresist, which is controlled by molecular size, and the affinity of solvents with photoresist, as shown in Fig. 2. The first step solvents can easily permeate into the crust because they have low molar volume (>60 cm³/mol), and the second step solvent which has higher affinity (RED < 1) dissolves the photoresist. AFM (Fig. 3) and XPS results show that there is no damage on the GaAs surface after the treatment with acetonitrile + DMSO two-step process.