@article{oai:ynu.repo.nii.ac.jp:00004192,
 author = {Tsuda, Hirotaka and Nakazaki, Nobuya and Takao, Yoshinori and Eriguchi, Koji and Ono, Kouichi},
 issue = {3},
 journal = {Journal of vacuum science ＆ technology. Second series. B, Microelectronics and nanometer structures, processing, measurement and phenomena},
 month = {May},
 note = {Atomic- or nanometer-scale surface roughening and rippling during Si etching in high-density Cl2 and Cl2/O2 plasmas have been investigated by developing a three-dimensional atomic-scale cellular model (ASCeM-3D), which is a 3D Monte Carlo-based simulation model for plasma-surface interactions and the feature profile evolution during plasma etching. The model took into account the behavior of Cl+ ions, Cl and O neutrals, and etch products and byproducts of SiClx and SiClxOy in microstructures and on feature surfaces therein. The surface chemistry and kinetics included surface chlorination, chemical etching, ion-enhanced etching, sputtering, surface oxidation, redeposition of etch products desorbed from feature surfaces being etched, and deposition of etch byproducts coming from the plasma. The model also took into account the ion reflection or scattering from feature surfaces on incidence and/or the ion penetration into substrates, along with geometrical shadowing of the feature and surface reemission of neutrals. The simulation domain was taken to consist of small cubic cells of atomic size, and the evolving interfaces were represented by removing Si atoms from and/or allocating them to the cells concerned. Calculations were performed for square substrates 50 nm on a side by varying the ion incidence angle onto substrate surfaces, typically with an incoming ion energy, ion flux, and neutral reactant-to-ion flux ratio of Ei = 100 eV, Γi 0 = 1.0×10 16 cm-2s-1, and Γn0/Γi0 = 100. Numerical results showed that nanoscale roughened surface features evolve with time during etching, depending markedly on ion incidence angle; in effect, at θi = 0°or normal incidence, concavo-convex features are formed randomly on surfaces. On the other hand, at increased θi = 45° or oblique incidence, ripple structures with a wavelength of the order of 15 nm are formed on surfaces perpendicularly to the direction of ion incidence; in contrast, at further increased θi≥75°or grazing incidence, small ripples or slitlike grooves with a wavelength of <5 nm are formed on surfaces parallel to the direction of ion incidence. Such surface roughening and rippling in response to ion incidence angle were also found to depend significantly on ion energy and incoming fluxes of neutral reactants, oxygen, and etch byproducts. Two-dimensional power spectral density analysis of the roughened feature surfaces simulated was employed in some cases to further characterize the lateral as well as vertical extent of the roughness. The authors discuss possible mechanisms responsible for the formation and evolution of the surface roughness and ripples during plasma etching, including stochastic roughening, local micromasking, and effects of ion reflection, surface temperature, and ion angular distribution. Moreover, plasma etching experiments of blank Si substrates in Cl2 were conducted by varying the rf bias power or ion incident energy to verify the validity of our ASCeM-3D model. A comparison of the etch rate and root-mean-square (rms) surface roughness between experiments and simulations indicated that the ASCeM-3D with θi = 0°reproduces well the experiments at Ei < 250 eV, while does not reproduce the rms roughness at higher Ei > 250 eV, where the roughness decreases with increasing Ei in experiments, while continues to increase with Ei in simulations. Possible mechanisms for this disagreement at increased Ei are discussed with the help of several plasma and surface diagnostics and classical molecular dynamics simulations for Si/Cl and Si/SiCl systems.},
 pages = {031212-1--031212-21},
 title = {Surface roughening and rippling during plasma etching of silicon: numerical investigations and a comparison with experiments},
 volume = {32},
 year = {2014}
}