Page 85 -
P. 85

3-40                                                             MEMS: Design and Fabrication


             medical evidence indicates that silicon even without protein adsorption control is benign in the body
             [Maluf, 2000]. In reality, it remains of considerable interest to improve the biocompatibility of silicon
             microdevices and engineer surfaces with reduced protein and cell adherence for the prevention of foul-
             ing and fibrotic response — especially when dealing with active silicon devices in contact with blood (e.g., a
             Si-based glucose sensor) [Zhang et al., 1998]. Even if significant control over protein adsorption and cell
             adhesion is achieved, Si remains a brittle material that may shard in the body and, just like glass, is to be
             avoided in any load-bearing application in living subjects or foodstuffs (e.g., as a micromachined needle).
             Although the motivation for most MEMS devices (especially active ones) is usually to eliminate any adhe-
             sion between proteins and surfaces, in some cases, the ability to enable protein adhesion locally but not
             in other areas and thus provide for a background signal (e.g., for cell-based assays) clearly requires two
             different surface chemistries. Studies at the University of Michigan and MIT, according to Issys’s Website,
             demonstrated that boron-doped single-crystal Si has a superior biocompatibility when compared with
             standard single-crystal Si or polysilicon-based devices. Some quite definitive work has been done with Si
             for cortical implants. Si electrode arrays were proposed as permanent cortical implants with the purpose
             of restoring useful vision, and their biocompatibility with neural tissue was investigated. In this context,
             silicon electrode arrays were left in place within a rabbit cortex for six months. Afterward, the neuron
             density as a function of radius from the center of the electrode shafts was used as the means of assessing
             damage. The researchers concluded that the electrode array showed long-term biocompatibility because
             normal brain tissue was observed within microns of the surface of the shafts [Clark et al., 1992; Horch
             et al., 1993].


             3.7    Wet Isotropic and Anisotropic Etching

             3.7.1    Introduction

             Wet etching is used for cleaning, shaping 3-D structures, removing surface damage, polishing, and char-
             acterizing structural and compositional features [Uhlir, 1956]. The materials etched include semiconductors,
             conductors, and insulators. The most important parameters in chemical etching are bias (undercut),
             tolerance, etch rate, anisotropy, selectivity, overetch, feature size control, and loading effects. The empha-
             sis in this chapter is on etching Si. Wet chemical etching of Si provides a higher degree of selectivity than
             dry etching techniques. Wet etching usually is also faster: a few microns to tens of microns per minute for
             isotropic etchants and about 1µm/min for anisotropic wet etchants vs. 0.1µm/min in typical dry etching.
             More recently, though, with inductively coupled plasma (ICP) dry etching of Si, rates of up to 6µm/min are
             achieved. Modification of wet etchant and/or temperature can alter the selectivity to silicon dopant con-
             centration and type and, especially when using alkaline etchants, to crystallographic orientation. Etching
             proceeds by (1) reactant transport to the surface, (2) surface reaction, and (3) reaction product transport
             away from the surface. If either (1) or (3) is rate determining, etching is diffusion limited and may be
             increased by stirring. If (2) is the rate-determining step, etching is reaction-rate limited and depends strongly
             on temperature, etching material, and solution composition. Diffusion-limited processes have lower acti-
             vation energies (on the order of a few Kcal/mol) than reaction-rate-controlled processes and therefore are
             relatively insensitive to temperature variations. In general, reaction-rate limitation is preferred, as it is eas-
             ier to reproduce a temperature setting than a stirring rate. The best generic etching apparatus has both a
             good temperature controller and a reliable stirring facility [Kaminsky, 1985; Stoller et al., 1970].
               Isotropic etchants (also polishing etchants) etch in all crystallographic directions at the same rate; they
             usually are acidic, such as HF/HNO /CH COOH (HNA), and lead to rounded isotropic features in single-
                                             3
                                                 3
             crystalline Si. The HNA etch is also known as the poly-etch because in the early days of the integrated cir-
             cuit industry it was used as an etchant for polysilicon [Maluf, 2000]. Isotropic etchants are used at room
             temperature or slightly above ( 50°C). Historically, they were the first Si etchants introduced [Robbins
             and Schwartz, 1959; 1960; Schwartz and Robbins, 1961; 1976; Uhlir, 1956; Hallas, 1971; Turner, 1958; Kern,
             1978; Klein and D’Stefan, 1962]. Later, it was discovered that some alkaline chemicals will etch anisotrop-
             ically; that is, they etch away crystalline silicon at different rates depending on the orientation of the



             © 2006 by Taylor & Francis Group, LLC
   80   81   82   83   84   85   86   87   88   89   90