Low-temperature plasma technology is used in as much as 70% of the front-end processes of semiconductor fabrication, including etching, deposition, and surface modification [1-7]. In this talk, I focus on the importance of controlling elementary processes in semiconductor process plasmas. First of all, the choice of source gas is important. In hard carbon thin film deposition, for instance, the higher the C/H in the film, the harder the film tends to be. Therefore, it is easier to achieve a higher degree of hardness by using C2H2 as the raw material gas rather than CH4, which has a lower C/H. If the raw material molecules contain benzene rings, the benzene rings are more easily incorporated into the films, and the C/H tends to be lower. Next, it is important to select the type of radicals produced by electron impact dissociation of the source gas; in the case of CH4, four types of radicals, CH3, CH2, CH, and C, are produced by electron impact dissociation. The surface sticking probability is small for CH3 and large for C. To deposit films with high step coverage on complexly shaped surfaces such as trenches, it is desirable to use CH3, which has a small surface sticking probability, as the main precursor for film deposition. These two different directions make it difficult to form hard carbon thin films in trenches with a high step coverage. Therefore, other avenues are to be explored. One such approach is control of elementary processes by unconventional plasma generation, which includes pulsed plasmas, amplitude-modulated RF discharge plasmas, and arbitrary waveform excitation plasmas. In short, elementary process control is the key to success in low-temperature plasma technology.Acknowledgements: This work was partly supported by JSPS KAKENHI (Grant Nos. JP20H00142, JP21H01372, JP21K18731, and JP23K03368), the JST ASPIRE, the NTT collaborative research, and the Murata Science Foundation.[1] M. Morimoto, et al., Jpn. J. Appl. Phys. 62 (2023) SN1001.[2] K. Kamataki, et al., Mater. Sci. Semicond. 164 (2023) 107613.[3] M. N. Agusutrisno, et al., Mater. Sci. Semicond. 162 (2023) 107503.[4] R. Narishige, et al., J. Mater. Res. 38 (2023) 1803.[5] N. Yamashita, et al., J. Mater. Res. 38 (2023) 1178.[6] I. Nagao, et al., Mater. Adv. 7 (2022) 918.[7] S. H. Hwang, et al., Processes., 9 (2020) 2.
Read full abstract