Miguel A. Gosalvez

Surface morphology during anisotropic wet chemical etching of crystalline silicon

The rich variety of micron-scale features observed in the orientation-dependent surface morphology of crystalline silicon during anisotropic wet chemical etching is shown to have its origin at the atomistic scale. Realistic Monte Carlo simulations show that the pyramidal hillocks on Si(100) are the result of local stabilization of distributed apex atoms by (metal) impurities from solution. In the absence of this stabilization, shallow round pits are formed on Si(100) as a result of the anisotropy between (one layer deep) pit nucleation and (isotropic) step propagation. It is also concluded that nosed zigzag structures at vicinal (110) are the combined result of misaligment and the etching anisotropy, showing that the nucleating mechanisms of morphologically related structures such as pyramidal hillocks and nosed zigzags are not necessarily the same. The simulations confirm that the formation of (one layer deep) triangular and hexagonal pits on exact Si(111) and of polygonal (saw-shaped) and straight terraces in vicinal Si(111) depends on the relative rate of [11] and [2] step propagation and on the misorientation of the surface with respect to Si(111).

Study of rounded concave and sharp edge convex corners undercutting in CMOS compatible anisotropic etchants

In this paper, we have studied the undercutting at rounded concave and sharp convex corners in (1 0 0)-silicon wafers using a complementary metal-oxide semiconductor (CMOS) compatible tetramethyl ammonium hydroxide (TMAH) solution with and without surfactant. In order to minimize the undercutting at both corner types while keeping reasonable etch rates, smooth etched-surfaces and CMOS compatibility, the non-ionic surfactant NC-200 that contains 100% polyoxyethylene-alkyl-phenyl-ether is considered. The effect of concentration and etching temperature is studied using 10, 20 and 25 wt% TMAH solutions at 60, 70 and 80 °C. When NC-200 at 0.1% of the total volume of the etchant is used, the undercutting ratio at both rounded concave and sharp convex corners is beneficially reduced as the etchant concentration is increased while, simultaneously, the etch rate increases. This is the opposite trend to the etch characteristics of pure TMAH. In addition, the rough etched surface morphology at low concentration is also improved by using NC-200.

Anisotropic wet chemical etching of crystalline silicon : atomistic Monte-Carlo simulations and experiments

Abstract An atomic-scale simulation model for anisotropic wet chemical etching of Si(1xa00xa00)-wafers covered with masks of arbitrary shape and a series of experiments for comparison are presented. The model assumes that the probability of removal of a surface atom depends on the number of first and second neighbors. A removal probability function is presented in order to describe the probabilities of removal corresponding to the different surface atoms having different numbers of first and second neighbors. Etching experiments on Si(1xa00xa00)-wafers using 10xa0wt.% KOH solution at 75°C are presented and the performance of the model is evaluated against them. We compare the underetched structures obtained in the simulations and experiments using a mask pattern consisting of a wagon wheel and a set of rectangular frame-like openings with varying orientation. The simulations show good agreement with the experiments. The model predicts the existence of fastest-etched planes in accordance with experiment, and describes accurately the evolution of under-etching below the masks for all mask orientations, including the slopes of the planes appearing below the mask. The results show that the cooperative effects of atoms evolving according to a simple rule generate most features of the meso- and macroscopic etching patterns. They also show that the use of only first neighbors or a partial incorporation of the second neighbors in the modeling strategy is not sufficient in order to describe the under-etching processes and that the second neighbors must be fully incorporated into the model.

Step flow-based cellular automaton for the simulation of anisotropic etching of complex MEMS structures

An enhanced octree data representation is developed for use in atomistic simulations of the evolution of complex multivalued surfaces appearing during anisotropic etching of crystalline silicon in micro electro mechanical systems (MEMS) applications. The octree provides a good balance between accuracy, memory efficiency and calculation speed. In combination with the octree, a step flow-based cellular automaton (CA) model is considered, which can be used to convert the experimental macroscopic etch rates into atomistic removal rates for direct use in the simulations. This involves minimizing the global error between the experimental values and the non-linear analytical expressions of the etch rates for a small set of chosen orientations. The orientations where the etch rate suffers a sudden change against the cutting angle play a crucial role to generate an optimal solution. The simulated etch rates for KOH and KOH/IPA systems in different concentrations show good agreement with the experiments. Due to the improved description of the anisotropy, the propagation of the etch front in realistic engineering applications can be simulated accurately.

Atomistic simulations of surface coverage effects in anisotropic wet chemical etching of crystalline silicon

Atomistic simulations of anisotropic wet chemical etching of crystalline silicon have been performed in order to determine the dependence of the etch rates of different crystallographic orientations on surface coverage and clustering of OH radicals. We show that the etch rate is a non-monotonic function of OH coverage and that there always exists a coverage value at which the etch rate reaches a maximum. The dependence of the anisotropy of the etching process on coverage, including the dependence of the fastest-etched plane orientation, is implicitly contained in the model and predictions of convex corner under-etching structures are made. We show that the whole etching process is controlled by only a few surface configurations involving a particular type of next-nearest neighbours. The relative value of the removal probabilities of these configurations determines the balance in the occurrence of step propagation and etch pitting for all surface orientations. # 2002 Elsevier Science B.V. All rights reserved.

Silicon Micromachining Based on Surfactant-Added Tetramethyl Ammonium Hydroxide: Etching Mechanism and Advanced Applications

This paper presents the mechanism behind the accused macroscopic changes in the etched profiles and etch rates caused by the addition of small amounts of surfactants (e.g., Triton X-100) in typical alkaline etchants (e.g., tetramethylammonium hydroxide or TMAH) for silicon micromachining applications targeting the fabrication of microelectromechanical systems (MEMS). In order to stress the technological importance of the surfactant addition in TMAH, the paper presents an overview of novel fabrication methods for the realization of new fixed and freestanding structures in Si{100} wafers using an inexpensive combination of wet anisotropic etching in pure and surfactant-added TMAH. The fixed structures contain perfectly sharp edges and a smooth etched surface finish. Thermally deposited oxide is used as the material for the freestanding structures. The fixed structures serve as molds for the realization of new structural shapes using poly(dimethylsiloxane) (PDMS).

Atomistic methods for the simulation of evolving surfaces

The currently available atomistic simulation tools for an evolving complex interface are described, focusing on the kinetic Monte Carlo (KMC) and cellular automata (CA) methods. The different versions of the two methods are considered, stressing their relative weaknesses and strengths. Data storage considerations lead to the choice of an octree data structure, of use in both KMC and CA simulations. The octree results in a substantial reduction in memory use, simultaneously benefiting from rapid addressability and fast searching properties. Examples in anisotropic etching are used to depict the use of the methods in practice and to describe the algorithms.

Effect of Cu impurities on wet etching of Si(110): formation of trapezoidal hillocks

We simulate the formation of experimentally observed trapezoidal hillocks on etched Si(110) surfaces, describing their generic geometrical shape and analyzing the relative stability and/or reactivity of the key surface sites. In our model, the hillocks are stabilized by Cu impurities in the etchant adsorbing on the surface and acting as pinning agents. A model of random adsorptions will not result in hillock formation since a single impurity is easily removed from the surface. Instead a whole cluster of Cu atoms is needed as a mask to stabilize a hillock. Therefore we propose and analyze mechanisms that drive correlated adsorptions and lead to stable Cu clusters.

Anisotropic Etching of Silicon as a Tool for Creating Injection Molding Tooling Surfaces

To improve the fidelity of the microinjection molding process, research is underway to implement silicon inserts as tooling surfaces in an injection molding machine. These tooling surfaces are created using typical microfabrication processes, such as bonding and chemical etching. The primary focus of this paper is the evaluation of anisotropic wet etching of off-axis silicon wafers, including experimental results and simulations. A method is presented for the determination of the lang110rang direction and subsequent alignment of the mask with an accuracy of 0.01deg. The use of atomistic kinetic Monte Carlo simulations reveals the extreme importance of proper alignment between the mask features and the off-axis wafer. As an example application, the fabrication steps and corresponding simulations of a silicon insert for the manufacture of disposable plastic razors are presented

Adsorption of metal impurities on H-terminated Si surfaces and their influence on the wet chemical etching of Si

We use first-principles methods to investigate the adsorption of Cu, Pb, Ag, and Mg onto a H-terminated Si surface. We show that Cu and Pb can adsorb strongly while Ag and Mg are fairly inert. In addition, adsorption states of two types are seen to exist for Pb. We also study the clustering energetics of Cu and Pb on the surface and find that while Cu clusters eagerly, Pb may prefer to form only small clusters of a few atoms. This kind of behavior of impurities is incorporated in kinetic Monte Carlo simulations of wet etching of Si. The simulation results agree with experiments supporting the idea that micromasking by Cu clusters and Pb atoms is the mechanism through which these impurities affect the etching process.