OES diagnostics as a universal technique to control the Si etching structures profile in ICP

Abstract

In this work, we demonstrate the high efficiency of optical emission spectroscopy to estimate the etching profile of silicon structures in SF6/C4F8/O2 plasma. The etching profile is evaluated as a ratio of the emission intensity of the oxygen line (778.1 nm) to the fluorine lines (685.8 nm and 703.9 nm). It was found that for the creation of directional structures with line sizes from 13 to 100 μm and aspect ratio from ≈ 0.15 to ≈ 5 the optimal intensities ratio is in the range of 2–6, and for structures from 400 to 4000 μm with aspect ratio from ≈ 0.03 to ≈ 0.37 it is in the range 1.5–2. Moreover, the influence of the process parameters on the etching rate of silicon, the etching rate of aluminum, the inclination angle of the profile wall of the etched window, the selectivity of silicon etching with respect to aluminum, and the influence on the overetching (Bowing effect) of the structure was investigated.

Chamber in-situ estimation during etching process by SiF4 monitoring using laser absorption spectroscopy

“…5(d)-5(f), the wavelength of 220 nm assigned to CF radical emission was used because 220 nm showed a significant change compared to the other wavelengths such as 252, 685, and 777 nm assigned to CF 2 , F, and O radical emission, respectively. [38][39][40][41] The intensity of the CF radical emission did not show a signal change as distinct as SiF 4 partial pressure, and the signal change was no longer detectable during monitoring of the 40% open area sample. This is presumably due to the etching condition where C 4 F 8 /O 2 gases were used as plasma sources.…”

Science-based, data-driven developments in plasma processing for material synthesis and device-integration technologies

“…313) The etching profiles of Si were estimated from the ratio of the intensities of the oxygen emission to the fluorine lines in the OES data for the SF 6 /C 4 F 8 /O 2 plasma. 314) To monitor the condition of the equipment and examine the potential cause of the fault, SVID was identified from sensor data using k-means and ML methods, neighbors (kNNs algorithms), 315) and naive Bayes classifiers. Multidimensional SVID are graphically depicted using tdistributed stochastic neighbor embedding.…”

In situ diagnostics of the Si etching structures profile in ICP SF6/C4F8 plasma: Macrostructures

“…Thus, in order to accelerate the development of etching technology, it is necessary to monitor the etching process in real time (in situ). Optical emission spectroscopy (OES) is the most suitable method to solve this kind of tasks 16,17 . In our previous work 18 , to create directional silicon structures with etching window diameter of 13 µm in SF 6 /C 4 F 8 /O 2 plasma, the O/F radical ratio was controlled using OES.…”