Fig 7 - uploaded by Josef Weinbub
Content may be subject to copyright.
![Comparison of simulated topography to scanning electron microscopy image from Panduranga et al. [4] for a trench with d = 8 μm. The undercut is shown by H and the etch depth by V. Reprinted with permission from J. Vac. Sci. Technol. B 37, 061206 (2019). Copyright 2019, American Vacuum Society.](profile/Josef-Weinbub/publication/358875972/figure/fig5/AS:1134498032951298@1647496817598/Comparison-of-simulated-topography-to-scanning-electron-microscopy-image-from-Panduranga.png)
Comparison of simulated topography to scanning electron microscopy image from Panduranga et al. [4] for a trench with d = 8 μm. The undercut is shown by H and the etch depth by V. Reprinted with permission from J. Vac. Sci. Technol. B 37, 061206 (2019). Copyright 2019, American Vacuum Society.
Source publication
![Fig. 4. Comparison of etch depths reported by Panduranga et al. [4]...](profile/Josef-Weinbub/publication/358875972/figure/fig2/AS:1134498032951296@1647496817535/Comparison-of-etch-depths-reported-by-Panduranga-et-al-4-with-simulation-results-in_Q320.jpg)
Low-bias etching of silicon (Si) using sulfur hexafluoride (SF6) plasma is a valuable tool in the manufacturing of electronic devices and micro electro-mechanical systems (MEMS). This kind of etching offers an almost isotropic etching behaviour, since the low voltage bias does not provide enough vertical acceleration and kinetic energy to the ions....
Contexts in source publication
Context 1
... geometries to the experimental data is shown in Fig. 6. We note that the reported experimental data exhibits noise due to the measurement accuracy, leading to challenges with regards to calibrating to the measured undercut. Additionally, a comparison of simulated to a scanning electron micrography image for a trench with d = 8 μm is shown in Fig. ...
Context 2
... Eq. (9), H refers to the horizontal etch depth or underetch, i.e., half the lateral extent of the trench at the top minus half the initial mask opening; and V refers to the vertical etch depth. These quantities are shown in Fig. 7. In ideally isotropic structures I = 1, while fully anisotropic vertical structures result in I = 0. This degree of isotropy has to be precisely controlled by the process recipe, even if the desired outcome is not necessarily I = 1. For example, for laser cavity applications, a process with I near 1 would result in a flat bottom which ...
Citations
... In the last three decades, the overwhelming research interest in the field of micro-electro-mechanical system (MEMS) for 'intelligent' sensors and devices can be attributed to two prime factors: integration of many novel materials with silicon [1][2][3][4][5][6][7], and the fast progress in the precision three-dimensional micromachining technologies [1,2,[8][9][10][11]. These MEMS-based 'intelligent' sensors and devices consist of a wide variety of J Mater Sci: Mater Electron (2023) 34:2270 2270 Page 2 of 10 footprint and overall cost; but also open the scope of adding more functionality by heterogenous integration of electronics, sensors, and photonics in the same chip (multi-stacking) [38][39][40]. ...
This paper presents etching optimization of comb structure formation in small silicon substrates mounted on a large silicon wafer (150 mm diameter) using the S1818 photoresist adhesive layer. A series of deep reactive ion etching (DRIE) experiments are done by varying the etching conditions and tweaking the baking schedule of the photoresist adhesive layer to get uniform vertical arrays of comb-fingers. The S1818 photoresist is also used as a masking layer (2.5 μm thick) for patterning a 16 μm deep vertical comb structure using the Bosch process-based DRIE. The optimized RF powers of etching (2200 W) and deposition (1900 W) cycles of the DRIE yielded minimal damage on the masking layer. The optimized soft-bake temperature (90 °C for 1 min) of the photoresist adhesive layer helps in achieving desired comb structure with smooth vertical sidewalls (scallops: <50 nm). The silicon etch rate in the optimized condition is found to be 3.5 (± 0.2) µm/ min with an etch selectivity of 20. The prime reason for achieving the smooth vertical comb-walls is the efficient RF power generated by heat dissipation through the moisture of the solvent (1-methoxy-2-propanol-acetate) present in the photoresist adhesive layer during the DRIE process. Finally, a bulk-micromachined comb-type MEMS accelerometer structure is successfully fabricated in a small silicon substrate after optimizing the smooth vertical comb-walls using DRIE. The fabricated accelerometer exhibited a 60 mV/g scale-factor sensitivity with 120 Hz 3dB bandwidth.
... where V ( x, t) stands for a velocity field that drives the evolution of the zero level-set. The modeling of the physical and chemical processes is achieved through the construction of V ( x, t) [8,9]. Certain velocity fields are so intricate that the properties of the Hamiltonian H = V ( x, t) ∇φ( x, t) in (2) change, e.g., H only has a weak solution [10]. ...
... Alongside with requiring a smaller stencil, the Shape Operator method has an additional performance advantage: It is possible to avoid explicitly calculating the Gaussian curvature to determine if a point on the zero level-set is a feature. Both conditions of the feature detection algorithm (8) are simultaneously checked if: ...
The level-set method is widely used in expanding front simulations in numerous fields of computational research, such as computer graphics, physics, or microelectronics. In the latter, the level-set method is employed for topography simulations of semiconductor device fabrication processes, being driven by complicated physical and chemical models. These models tend to produce surfaces with critical points where accuracy is paramount. To efficiently increase the accuracy in regions neighboring these critical points, automatic hierarchical domain refinement is required, guided by robust feature detection. Feature detection has to be computationally efficient and sufficiently accurate to reliably detect the critical points. To that end, we present a fast parallel geometric feature detection algorithm for three-dimensional level-set functions. Our approach is based on two different, complementary curvature calculation methods of the zero level-set and an optimized feature detection parameter to detect features. For performance reasons, our algorithm can be in principal linked to different curvature calculation methods, however, as will be discussed, two particularly attractive options are available: (i) A novel extension of the standard curvature calculation method for level-set functions, and (ii) an often disregarded method for calculating the curvature due to its purported low numerical accuracy. We show, however, that the latter is still a viable option, and that our algorithm is able to reliably detect features on geometries stemming from complicated, practically relevant geometries. Our algorithm and findings are applicable to other fields of applications such as surface simplification.
