Abstract

Benzocyclobutene (BCB) has been widely used as passivation and interlayer dielectric for sub-millimeter-wave applications and terahertz monolithic microwave integrated circuits (TMICs). In this work, the influence of BCB passivation on 100-nm InP-based high-electron-mobility transistors (HEMTs) was investigated. A set of 100-nm HEMTs with different gate recess structures were fabricated and compared. A SiO <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$_{\text{2}}$</tex-math> </inline-formula> hard mask was adopted in the gate recess, which contributed to an improved device performance after BCB passivation. DC and RF performances were characterized before and after BCB passivation. The results show that BCB passivation in the recess region degraded InP HEMTs’ performance, and a better performance was achieved by the SiO <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$_{\text{2}}$</tex-math> </inline-formula> hard mask. For traditional devices, of which the gate recess was exposed to silicon nitride (SiN <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$_{\textit{x}}\text{)}$</tex-math> </inline-formula> and BCB passivation layers, both dc and RF performances were deteriorated drastically even though the thickness of SiN <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$_{\textit{x}}$</tex-math> </inline-formula> passivation layer was increased from 20 to 50 nm. However, the SiO <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$_{\text{2}}$</tex-math> </inline-formula> hard mask preserved the device performances to a great degree, with only degradation by increased parasitic capacitances according to the small-signal equivalent model. In addition, the threshold voltage shift after BCB passivation was suppressed by the SiO <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$_{\text{2}}$</tex-math> </inline-formula> hard mask. Thus, a device structure suitable for BCB passivated TMIC applications is proposed and validated.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call