蚀刻(微加工)
分析化学(期刊)
氢
等离子体刻蚀
原子层沉积
傅里叶变换红外光谱
图层(电子)
沉积(地质)
化学
材料科学
纳米技术
化学工程
有机化学
古生物学
工程类
生物
沉积物
作者
Shih‐Nan Hsiao,Makoto Sekine,Masaru Hori
标识
DOI:10.1021/acsami.3c04705
摘要
Cyclic atomic layer etching (ALE) of SiN with high selectivity to SiO2, utilizing a hydrofluorocarbon deposition followed by exposure to hydrogen plasma, is presented. The surface reaction mechanism and etching behavior were investigated with in situ attenuated total reflectance Fourier transformation infrared spectroscopy (ATR-FTIR) and spectroscopic ellipsometry. In the deposition step, the hydrofluorocarbon film was deposited on top of the SiN films using the CF4/H2 plasmas with varying H2 contents (33 to 85%). Subsequently, the surface-modified SiN film was exposed to a hydrogen plasma for etching. The self-limiting SiN etching was observed, where the etch depth solely depended on the F concentration of the deposited hydrofluorocarbon layer once its thickness exceeded a critical value. A high selectivity of approximately 8.6 for SiN over SiO2 was achieved. The in situ ATR-FTIR spectra revealed that during the deposition step, besides the formation of the C-H peak associated with hydrofluorocarbon deposition, the appearance of the N-H4 absorbance band indicated the formation of an ammonium fluorosilicate layer on top of SiN. In the subsequent H2 plasma etching step, both the surface modification layer and the pre-deposited hydrofluorocarbon layer were removed. The removal of the surface-modified layer and hydrofluorocarbon layer was associated with the etch rate during H2 plasma exposure. These findings indicate the importance of the formation and removal of the surface modification layer for achieving ALE of SiN. The dissociation of the hydrofluorocarbon layer by the H2 plasma released reactants that interacted with SiN, leading to the formation of a new surface modification layer. The etching process significantly slowed down once the hydrofluorocarbon deposition and surface modification layer were completely removed.
科研通智能强力驱动
Strongly Powered by AbleSci AI