Wet Etch vs Dry Etch

Etching removes material that lithography has already exposed, and a fab chooses between two fundamentally different ways to do it: wet etch dissolves material with liquid chemicals, while dry etch blasts it away with a reactive plasma. The choice matters because each method removes material in a different shape, and the shape of that removal decides whether a modern chip's microscopic features come out usable or ruined.

Wet etch removes material in every direction at once, which is exactly why it can't pattern today's smallest features. Dipping a wafer in a liquid etchant like hydrofluoric acid dissolves the exposed material uniformly, eating sideways under the photoresist mask just as fast as it eats downward — a property called isotropic etching. That sideways bite, called undercut, is harmless when features are relatively large, but at nanometer-scale dimensions it destroys the precise, vertical-walled shapes a modern transistor or interconnect line needs.

Dry etch solves the undercut problem by making the removal directional. A plasma chamber accelerates reactive ions toward the wafer largely perpendicular to its surface, so material gets bombarded away where the ions land and is mostly protected wherever the photoresist mask covers it — this directionality is called anisotropic etching, and it is what allows a fab to cut a nearly vertical wall straight down into a film only tens of nanometers wide.

Neither method has fully replaced the other, because they solve different problems well. Dry etch dominates the critical, feature-defining steps at advanced nodes — gate patterning, contact holes, fine interconnect lines — precisely because those steps cannot tolerate any undercut. Wet etch remains the workhorse for less dimensionally critical jobs: cleaning native oxide off a wafer before a process step, stripping a sacrificial layer, or removing blanket films where a little sideways etching does no harm and a liquid bath is simply cheaper and gentler on the wafer.

Selectivity is the other half of the etch decision, alongside directionality. A good etch recipe removes the target film quickly while barely touching the layers above, below, or beside it — a plasma etch tuned for silicon dioxide, for instance, needs to stop cleanly once it hits an underlying silicon layer rather than continuing to dig in. Both wet and dry chemistries are engineered around this selectivity ratio, because an etch that is perfectly directional but poorly selective can still ruin a device by eating into a layer it was never supposed to touch.

Read etch selection through a directionality lens: the question is never simply "which method removes material faster or cheaper," it is "does this specific step need a vertical wall or can it tolerate a wider, undercut shape" — and that answer, more than cost or chemistry, is what decides whether a given process step at a modern fab runs wet or dry.