M.Sc Thesis

M.Sc StudentNachmias Tali
SubjectDeep Dry Etch (DRIE) Processes in Silicon for 3-D
DepartmentDepartment of Chemistry
Supervisors PROFESSOR EMERITUS Chaim Yarnitzky (Deceased)


The major research subject of this thesis is deep reactive ion etching of Silicon in DRIE machine. The BOSCH process enables 90º sidewall angles structures with high aspect ratio. One aim of this research is to develop an improved BOSCH process, which characterized by its smoother scalloping sidewall (roughness) comparing to the original Bosch process, by reducing the etching time. The roughness obtained from the new process is: lu value (peak to valley-undercut) is assumed to be 0nm and ld value (depth per cycle) is measured to be 300nm.

The second aim of this research is the improvement of negatively tapered sidewall angle of large trenches using sacrificial beams.

We observed 2 factors which improve the sidewall angle:                                           

1. Placing the sacrificial beam closer to the sidewall improves the sidewall angle towards the 90°. The explanation for this phenomenon is that ions attract to negatively conductive trench sidewalls. Because of that, ion which enter to a narrow trench will be less attracted to its closest sidewall because the opposite sidewall will apply an opposite attraction force. Consequently the ion will not feel any attraction forces, it trajectory will remain vertical and it will hit the bottom of the trench. 2. Placing several sacrificial beams (more then 1 sacrificial beam), will cause 0.5-1.5° improvement in the sidewall angle in the transition from 1 sacrificial beam design to 3 sacrificial beams design. This can be regarded to the micro-loading effect.

The last part of this research dealt with the questions: Are the beam undercut as a consequence of etching, influenced by the beam width and the trench width? The results show that the beam undercut is not dependent on both: the beam width and the trench width. The conclusion from this result is that the first assumption that we made which claims that ion flow is bigger for width trench than narrow trench is false. This assumption was made under the assumption that width trenches are exposed to larger ion angles than narrow trenches. Due to that, width trenches are exposed to large ion flow, which probably causes to larger undercut rather than in the narrow trenches. Since the results are not supporting this assumption and the beam undercut is identical for width and narrow trenches, that means that the ion flows for both trenches are identical.