The TFE process uses three masks and a variety of masking levels to pattern and etch the wafer layers. The process uses double-side polished, fused silica wafers that measure 6 in and 500 µm thick. Diagrams below are not shown to scale, sizes have been exaggerated for educational purposes.
A 5 µm layer of polyimide is deposited onto the sapphire substrate via spin coating. The wafer is then cured in nitrogen gas for 30 minutes at 350°C.
Two layers of resist are applied to the top of the wafer.
The photoresist layers are patterned with the light field METAL1 mask.
A 240 nm layer of pad metal is deposited on top of the wafer—this is composed of 20 nm Ti, 20 nm Pt, 160 nm Au, 20 nm Pt, and 20 nm Ti.
The photoresist is removed with a chemical strip.
A new layer of polyimide is applied across the top of the wafer.
Two layers of resist are applied in the same pattern as the METAL1 mask.
A thin layer of metal is applied to the top of the wafer as a hard mask.
The two layers of resist are removed in a chemical strip.
The patterned metal hard mask guides the etch of the polyimide 2 layer while the OUTLINE1 masking level guides an etch through the full thickness of the wafer.
A second etch removes the hard mask and the topmost layer of the pad metal stack.
Two layers of resist are applied across the top of the wafer.
The photoresist is etched with the TOPMETAL mask.
The 100 nm top metal layer is applied across the top of the wafer—this is composed of 10 nm of Ti and 90 nm of Pt. The portion of the top metal layer that extends past the VIA masking level may not be electrically connected to the metal 1 layer.
The layers of resist are removed with a chemical etch.
A laser release removes the bottom substrate from under the area according to the RELEASE masking level.