A generalized model describing corner undercutting by the experimental analysis of TMAH/IPA

Up to now the etch fronts developed in anisotropic wet chemical etching by undercutting at convex corners have been assumed to be straight and correlated to some specially defined fast etching crystal planes. In particular, the propagation direction has been proposed to be in coincidence with the direction of maximal etch rate. These were found to be in accordance with the experimental data far potassium hydroxide. However, when we use tetramethylammonium hydroxide/isopropyl alcohol as the etchant this usual idea leads to severe contradictions with the experimental data, which reveal on the one hand a curved etch front and on the other hand an obvious discrepancy of the propagation direction from the direction of maximal etch rate. Further investigations also show the same deviation for TMAH and KOH/IPA. To overcome this dilemma we propose a generalized model considering all crystal planes to be equivalent at first and then determining the etch front curve by the projected etch rates.