Leakage Current Control in Recessed SiGe Source/Drain Junctions

Abstract

The impact of different process parameters, namely, the trench etch depth, the total epitaxial SiGe thickness, and the epi elevation, on the leakage current of recessed Si 0.8 Ge 0.2 source/drain junctions has been systematically investigated. Besides the behavior of the forward and the reverse currents, attention is also given to the temperature dependence of the leakage current. It is found that both the bulk and the peripheral leakage current density increase strongly with increasing etch depth. Empirically, an exponential dependence has been observed between the area leakage current density at Ϫ1 V and the distance d j between the Si 0.8 Ge 0.2 -Si interface and the electrical p-n junction, whereby an increase by 1 dec for every 43 nm of reduction in d j occurs. This can be understood by the fact that the responsible defects originate mainly at the SiGe-Si interface. The perimeter current density shows for certain process splits an exponential dependence on the total thickness of the epitaxial layer t SiGe , with an increase by a decade for every 50 nm increase in thickness. Also, the generation and recombination lifetimes have been studied in order to determine an effective energy level of the electrically active defects.

Influence of the strain‐relaxation induced defect creation on the leakage current of embedded Si1–xGex source/drain junctions

“…This fact indicates that not only the diffusion mechanism contributes to the leakage, but also the SRH mechanism shows an influence in the high temperature range. This is in agreement with previous results [6], where lower activation energies than the Si bandgap were obtained for SiGe S/D junctions at high temperature regime. On the other hand, as is anticipated the activation energies decrease with the bias, due to the presence of high electric fields in the silicon depletion region and fieldassisted generation current mechanisms.…”

“…The effect of the diffusion component at room temperature is usually small and can be neglected [6]. Therefore, the reverse current density will be mainly dominated by the mid-gap-level generation processes in the depletion region, for not too shallow junctions and at 300 K. According to the Shockley-Read-Hall (SRH) theory, the generation current density is given by [3][4][5][6][7]:…”

“…The perimeter (J P ) and area (J A ) leakage current components have been extracted at fixed reverse bias from the slope and intercept of the reverse current density J R versus P/A ratio plot, given by (1) [6][7][8]. This approach is valid when corner and parasitic components can be neglected.…”

“…Additionally, C-V measurements at 1 MHz were used to derive the depletion width and the free carrier doping profiles in the Si n-well region. Finally, I-V measurements have been carried out between 25 and 200 °C in order to extract the activation energy of the leakage current components [6,9]. The presence of dislocations was evaluated by Nomarski microscopy and HRTEM.…”

Modeling impact of electric field and strain on the leakage of embedded SiGe source/drain junctions

“…One of the methods is by determining the activation energies (E A ) from a temperaturedependence study of current-voltage (I-V) curves [5][6]. Around room temperature, the main leakage mechanisms are SRH, TAT, and BTBT, while for devices working at sufficiently high temperatures the effect of the diffusion component [7] is becoming dominant. These temperaturedependence studies can also determine the ranges of reversebiased voltages where one of the previous mechanisms is dominant [8].…”