Performance, scalability and resilience to variability of Si SOI FinFETs and gate-all-around (GAA) nanowires (NWs) are studied using in-house-built 3D simulation tools. Two experimentally based devices, a 25 nm gate length FinFET and a 22 nm GAA NW are modelled and then scaled down to 10.7 and 10 nm gate lengths, respectively. A TiN metal gate work-function granularity (MGG) and line edge roughness (LER) induced variability affecting OFF and ON characteristics are investigated and compared. In the OFF-region, the FinFETs have over an order of magnitude larger OFF-current that those of the equivalent GAA NWs. In the ON-region, the 25/10.7 nm gate length FinFETs deliver 20/58% larger ON-current than the 22/10 nm gate length GAA NWs. The FinFETs are more resilient to the MGG and LER variability in the sub-threshold compared to the GAA NWs. However, the MGG ON-current variability is larger for the 10.7 nm FinFET than that for the 10 nm GAA NW. The LER ON-current variability depends largely on the RMS height; whereas a 0.6 nm RMS height yields a similar variability for both FinFETs and GAA NWs. Finally, the industry preferred 110 channel orientation is more resilient to the MGG and LER variability in both architectures.
Nanosheet (NS) and nanowire (NW) FET architectures scaled to a gate length (LG) of 16 nm and below are benchmarked against equivalent FinFETs. The device performance is predicted using a 3D finite element drift-diffusion/Monte Carlo simulation toolbox with integrated 2D Schrödinger equation based quantum corrections. The NS FET is a viable replacement for the FinFET in high performance (HP) applications when scaled down to LG of 16 nm offering a larger on-current (ION) and slightly better sub-threshold characteristics. Below LG of 16 nm, the NW FET becomes the most promising architecture offering an almost ideal sub-threshold swing, the smallest off-current (IOFF), and the largest ION/IOFF ratio out of the three architectures. However, the NW FET suffers from early ION saturation with the increasing gate bias that can be tackled by minimizing interface roughness and/or by optimisation of a doping profile in the device body.
The effect of gate edge roughness (GER) in the sub-threshold region is studied for two state-of-the-art architectures: a 10.7-nm Si FinFET and a 10-nm Si gateall-around (GAA) nanowire (NW) FET using an in-house 3D quantum-corrected drift-diffusion simulation tool. The GER is applied to the device gate using the characteristic values of root-mean-square amplitude and correlation length (CL). The GER-induced variability results in a standard deviation (σ) for the threshold voltage (V T ) of 7 mV for the FinFET when CL/Gate Perimeter = 0.66 and RMS = 0.80 nm, which is 20% greater than that of the GAA NW FET. GER is a less damaging source of variability than metal grain granularity (MGG), line edge roughness (LER), and random dopants (RD) for both devices. When compared to LER variations, σV T due to the GER is 62% and 86% lower for the FinFET and GAA NW FET, respectively. However, although GER affects the FinFET more than the GAA NW FET, the combined variability effect of GER, MGG, LER, and RD (σV T,comb ) on the FinFET is 30 mV, a value approximately 50% smaller than that of the GAA NW FET.
An in-house-built three-dimensional multi-method semi-classical/classical toolbox has been developed to characterise the performance, scalability, and variability of state-of-the-art semiconductor devices. To demonstrate capabilities of the toolbox, a 10 nm gate length Si gate-all-around field-effect transistor is selected as a benchmark device. The device exhibits an off-current ( ) of A/ m, and an on-current ( ) of 1770 A/ m, with the ratio , a value larger than that of a nm gate length Si FinFET. The device SS is 71 mV/dec, no far from the ideal limit of 60 mV/dec. The threshold voltage standard deviation due to statistical combination of four sources of variability (line- and gate-edge roughness, metal grain granularity, and random dopants) is mV, a value noticeably larger than that of the equivalent FinFET (30 mV). Finally, using a fluctuation sensitivity map, we establish which regions of the device are the most sensitive to the line-edge roughness and the metal grain granularity variability effects. The on-current of the device is strongly affected by any line-edge roughness taking place near the source-gate junction or by metal grains localised between the middle of the gate and the proximity of the gate-source junction.
2 November 2017 in IEEE Transactions on Electron Devices
The impact of the metal grain granularity (MGG) variations on subthreshold and ON-current of a 22-nm gate length Si gate-all-around (GAA) nanowire (NW) field-effect transistor (FET) is analyzed by assessing five figures-of-merit: threshold voltage ( ${V}_{T}$ ), OFF-current ( ${I}_{ \mathrm{\scriptscriptstyle OFF}}$ ), subthreshold slope (SS) and drain-induced-barrier-lowering at low ( ${V}_{D}={50}$ mV) and high ( ${V}_{D}={1.0}$ V) drain biases, and ON-current variability at the high drain bias. We assume a TiN metal gate with four grain sizes (GS): 10, 7, 5, and 3 nm. The simulations carried out employ: 1) a 3-D finite element (FE) drift-diffusion (DD) simulator with density-gradient (DG) quantum corrections (QC) in the subthreshold region and 2) a 3-D FE Monte Carlo (MC) simulation toolbox with Schrödinger equation QC for the ON-current. We have found that the ${\sigma }{V}_{T}$ due to the MGG variability is 10% higher in the GAA-NW FET (GS of 10 nm) than in a 20-nm gate FinFET. For a GS of 3 nm, the variability becomes comparable between the GAA-NW FET and the FinFET. Comparison against the line-edge-roughness (LER) variability shows that the MGG (GS of 10 nm) affects ${\sigma }{V}_{T}~92$ % more than the LER (root-mean-square height of 0.85nm and correlation length of 20 nm). Finally, the ON-current of the FinFET (GS of 10 nm, ${V}_{D}=\textsf {0.9}$ V) has a 43% higher ${\sigma }{I}_{ \mathrm{\scriptscriptstyle ON}}$ than that of the GAA-NW (GS of 10 nm, ${V}_{D}=\textsf {1.0}$ V) making the GAA-NW FETs a much better solution for sub-10-nm technology nodes.
6 May 2016 in IEEE Transactions on Electron Devices
We have demonstrated, via validation to experimental data for TiN and Ru, that the grains that appear in the metal gate-stacks of nanoscale CMOS devices can be characterized via a two-parameter Gamma distribution (p -values 0.17 and 0.42 for TiN and Ru). Conversely, a previously presented fit that used Rayleigh distribution does not reproduce the experimental data (p -values 3 × 10 -14 and 0.0029 for TiN and Ru). Poisson Voronoi diagrams (PVDs) are shown as a suitable algorithm to generate grains with Gamma distribution, via fitting of the distribution of 10 000 grains. We have also compared the PVD variability against the Rayleigh model for both TiN and TaN metal gates, and concluded that Rayleigh approach overestimates the device variability (by 11.9% for the TiN and by 7.14% for the TaN).