P.-C. ChiuY.-C. LuVITA PI-HO HU2020-10-072020-10-07201921686734https://scholars.lib.ntu.edu.tw/handle/123456789/516584https://www.scopus.com/inward/record.uri?eid=2-s2.0-85062696889&doi=10.1109%2fJEDS.2019.2897286&partnerID=40&md5=aab698eb7cce668f5d74e1da1b0ab575In this paper, the impacts of work function variation (WFV), line-edge roughness (LER), and ferroelectric properties variation on the threshold voltage, subthreshold swing (SS), Ion, and Ioff variations are analyzed comprehensively for negative capacitance ultra-thin body SOI MOSFETs (NCSOI) compared with SOI MOSFETs (SOI). For LER induced threshold voltage variation ( σ Vt), NC-SOI MOSFETs exhibit smaller σ Vt (= 3.8 mV) than the SOI MOSFETs ( σ Vt = 17.6 mV). For analyzing WFV of NC-SOI MOSFETs, two scenarios are considered including (I) same WFV patterns, and (II) different WFV patterns between the external and internal metal gates. Compared with SOI, NC-SOI with scenario (I) exhibits comparable WFV induced σ Vt (= 16.2 mV), and NC-SOI with scenario (II) exhibits larger WFV induced σ Vt (= 28.5 mV). In scenario (II), different WFV patterns between the internal and external gates result in VFE (voltage drop across the ferroelectric layer) variations, which increases the WFV induced σ Vt for NC-SOI. LER dominates energy-delay product variations ( σ EDP), and NC-SOI MOSFETs show smaller σ EDP than SOI MOSFETs. Besides, NC-SOI MOSFETs with thicker ferroelectric layer thickness ( TFE ), larger coercive electric field ( EC), and smaller remnant polarization (P0) show smaller LER induced σ Vt and σ SS. Ferroelectric properties variations show negligible impact on the WFV induced σ Vt and σ SS. © 2013 IEEE.ferroelectric properties; Line-edge roughness; negative capacitance FET (NCFET); work function variationCapacitance; Electric fields; Ferroelectricity; MOSFET devices; Threshold voltage; Work function; Coercive electric field; Ferroelectric property; Function variation; Line Edge Roughness; Negative capacitance; Remnant polarizations; Sub-threshold swing(ss); Threshold voltage variation; Roughness measurementjournal article10.1109/jeds.2019.28972862-s2.0-85062696889