Sacrificial Layer

A sacrificial layer is a temporary material deliberately deposited during MEMS fabrication specifically so it can be removed later, freeing the movable structures built on top of or around it and defining the air gaps and mechanical clearance that let those structures actually move. It's a genuinely counterintuitive design concept relative to conventional semiconductor processing, where every deposited layer is normally meant to stay — here, an entire deposition step exists purely to occupy space temporarily, with its removal being just as deliberately engineered as its initial deposition.

Etch selectivity between the sacrificial material and the surrounding structural and anchor materials is the single physical property this entire fabrication strategy depends on, because the release etch step has to remove the sacrificial layer essentially completely while leaving every structural film and anchor point it contacts along the way almost entirely untouched. This selectivity requirement is exactly why sacrificial-structural material pairing isn't an arbitrary choice — a sacrificial material and structural material only work well together in a surface-micromachined MEMS process if a genuinely selective etchant exists that attacks one aggressively while leaving the other essentially intact, and that specific chemistry pairing has to be established and validated before the process can actually work.

Release hole placement, positioning small access openings specifically to let etchant physically reach and penetrate the sacrificial layer throughout its entire extent, directly determines whether release actually completes successfully, because an etchant that can't physically reach some interior region of the sacrificial layer simply leaves that region unremoved regardless of how long the etch step runs. This access-design requirement is exactly why release-hole placement is treated as a genuine, deliberate layout design decision rather than an afterthought — a sacrificial-layer geometry without adequate etchant access points can leave the finished device with trapped, unreleased material no matter how aggressive or prolonged the etch chemistry itself is.

Sacrificial layer thickness directly and precisely sets the resulting air-gap height and mechanical motion range once release is complete, which means the deposition step creating the sacrificial layer has to control thickness with the same precision that device geometry itself ultimately depends on. This direct thickness-to-gap-height relationship is exactly why sacrificial layer deposition can't be treated as a rough, uncritical process step — the final device's mechanical behavior, capacitance, and motion range are all set by a dimension that exists only temporarily during fabrication and gets entirely removed before the finished device ever operates.

Stiction — unwanted adhesion between the released structure and the substrate or neighboring structures, typically arising from capillary forces during wet release-etch rinsing and drying — represents one of the most common and consequential yield-loss mechanisms specific to sacrificial-layer-based MEMS fabrication, and it's exactly why controlled rinsing and drying procedures receive dedicated process attention after release. This stiction risk is a direct consequence of the release process's own wet-chemistry nature — the same liquid etchant and rinse steps that successfully remove the sacrificial material can, if not carefully controlled during the drying step, pull the newly freed structure into permanent contact with an adjacent surface through capillary force, converting a successfully released structure into a stuck, non-functional one.

Sacrificial-layer residue and etch byproducts left behind after an incomplete or imperfectly controlled release represent a frequent, specific failure source distinct from stiction, because residual material can mechanically obstruct motion or electrically interfere with device operation even when the structure hasn't actually adhered to anything. This residue-related failure mode is exactly why post-release cleaning gets treated as a genuinely critical process step rather than a routine cleanup formality — a release etch that successfully removes the bulk sacrificial material while leaving behind enough residue to interfere with device function has only partially succeeded, and distinguishing this residue-related failure mode from stiction matters directly for correctly diagnosing why a given released device isn't working as intended.