Deep silicon etching: STiGer vs. BOSCH - Archive ouverte HAL Accéder directement au contenu
Poster De Conférence Année : 2016

Deep silicon etching: STiGer vs. BOSCH

Résumé

Silicon deep etching is widely used in microelectronics and micro-electro-mechanical systems (MEMS). To achieve deep etching, cryogenic and Bosch processes are commonly used. The STiGer process is a cryogenic process introduced by GREMI and STMicroelectronics few years ago [1-3]. It consists in alternating SF6 plasma etching steps with SiF4/O2 plasma passivation steps. It is quite similar to the Bosch process, but it does not use fluorocarbon based species and only works at a very low temperature. The fact that a very low temperature is needed can appear as a drawback, but actually, it can be a significant advantage in terms of contamination. In the STiGer process, the passivation layer only forms on cooled surfaces. Hence, nearly no deposition is obtained on the reactor wall with the STiGer process so that cleaning steps are not necessary after etching, unlike with the Bosch process. In this paper, we compare the etching performances achieved by Bosch and by STiGer processes. We will focus on etch rate, etching defects, selectivity and robustness. (100) silicon samples were partially covered by a mask of SiO2. In table 1, process parameters are given for both processes after 30 cycles of etching and passivation. Etching step duration was kept the same for both processes. We had to adjust the passivation step time to guarantee comparable profiles. In figure 1, SEM images of the etching profile of a 4 μm wide trench obtained by the two processes are shown. First, the trench sidewall is quite vertical for both processes (slightly more negative with the STiGer process). The etched depth was respectively as high as 27.8 μm using STiGer and 31.23 μm using the Bosch process (see table 2). However, the total duration of the etching process was lower using the STiGer process due to a lower passivation time leading to an equivalent averaged etch rate of 2.5 μm.min-1 in these conditions. Note that the etch rate was not optimized in this particular study. The objective was just to evaluate the two processes in nearly the same conditions of flow, pressure and power. A higher selectivity was obtained using the STiGer process. The scalloping seems to be more pronounced using the STiGer process. Measurements were carried out for other pattern dimensions and geometries. Detailed results will be presented at the workshop. As a conclusion of this study, the STiGer process is a robust cryogenic process quite equivalent to the Bosch process leading to higher throughput in terms of units per hour since no cleaning step is required between each wafer batch. Bosch STiGer Etching Passivation Etching Passivation References 1. Tillocher T, Dussart R, Overzet L J, Mellhaoui X, Lefaucheux P, Boufnichel M and Ranson P 2008 Two cryogenic processes involving SF6, O2, and SiF4 for deep etching J. Electrochem. Soc. 155 D187 2. Dussart R, Tillocher T, Lefaucheux P, Ranson P, Mellhaoui X, Boufnichel M and Overzet L J 2008 Deep anisotropic silicon etch method France Patent 2914782-A1 (assigned to STMicroelectronics, CNRS and University of Orleans) 3. Dussart R, Tillocher T, Lefaucheux P and Boufnichel M 2014 J. Phys. D: Appl. Phys. 47 123001
Fichier non déposé

Dates et versions

hal-01344704 , version 1 (12-07-2016)

Identifiants

  • HAL Id : hal-01344704 , version 1

Citer

Chengkun Tu, Ying Cui, Remi Dussart, Thomas Tillocher, Philippe Lefaucheux. Deep silicon etching: STiGer vs. BOSCH. Plasma Etch and Strip in Microtechnology, May 2016, Grenoble, France. ⟨hal-01344704⟩
287 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More