Bulk vs. Surface Micromachining

Bulk micromachining uses the substrate itself as the building material: deep etching into the wafer exposes three-dimensional structures such as membranes, cavities, or trenches. The anisotropic KOH/TMAH etching covered in an earlier chapter is a classic example – as is the DRIE process, which enables even steeper, deeper structures. Because this works directly in the crystal, bulk micromachining can produce comparatively large, mechanically robust structures, often tens to hundreds of micrometers deep.

The drawback: bulk micromachining consumes wafer volume and can only be combined to a limited extent with classical CMOS electronics on the same chip, since the deep etch steps could damage sensitive circuit structures. In addition, the crystal orientation of the substrate largely determines the possible shapes – on (100) silicon, for example, always the characteristic 54.74° angle, unless DRIE is used instead.

Surface micromachining takes the opposite approach: instead of etching into the substrate, thin layers are built up on the wafer surface and then partially removed – exactly the principle of the sacrificial layer technique covered in an earlier chapter. The actual mechanical structure is formed from deposited layers just a few micrometers thick, such as polysilicon, rather than from the substrate itself.

This allows significantly smaller, more finely structured devices and tighter integration with classical CMOS electronics on the same chip, since fabrication uses process temperatures and tools similar to standard semiconductor manufacturing. The trade-off: the resulting structures are mechanically less robust than those from bulk micromachining, since the thin layers withstand lower forces before breaking or permanently deforming.