Tantalum etch

Initially we etched tantalum using a reactive-ion etch (8:3:2 CHF₃:SF₆:Ar chemistry at 50 mTorr, RF/ICP power of 100/100 W). However, scanning electron microscopy (SEM) images showed that reactive-ion etches can produce rough edges as well as small pillars and boulders near the sidewalls, likely due to micromasking (Supplementary Fig. ²a, b). The anomalous objects in Supplementary Fig. ²b remained after the device was cleaned in piranha solution and treated in an oxygen plasma. In order to avoid these fabrication problems, we employed a wet etch composed of 1:1:1 HF:HNO₃:H₂O. We found that several resists delaminated before the tantalum was etched through, leaving the sidewalls and nearby tantalum visibly rough in SEM (Supplementary Fig. ²c). This problem was circumvented by using thick AZ 1518 resist (~2-μm tall), which left cleaner sidewalls (Supplementary Fig. ²d). Comparing Devices 4–10 with Devices 11–18 in Supplementary Table ¹, we note that the optimized wet etch likely improved T₁.