The through silicon via （TSV） technology has proven to be the critical enabler to realize a three-dimensional （3D） gigscale system with higher performance but shorter interconnect length. However, the received digital signal after trans- mission through a TSV channel, composed of redistribution layers （RDLs）, TSVs, and bumps, is degraded at a high data-rate due to the non-idealities of the channel. We propose the Chebyshev multisection transformers to reduce the signal reflec- tion of TSV channel when operating frequency goes up to 20 GHz, by which signal reflection coefficient （$11） and signal transmission coefficient （$21） are improved remarkably by 150% and 73.3%, respectively. Both the time delay and power dissipation are also reduced by 4% and 13.3%, respectively. The resistance-inductance-conductance-capacitance （RLGC） elements of the TSV channel are iterated from scattering （S）-parameters, and the proposed method of weakening the signal reflection is verified using high frequency simulator structure （HFSS） simulation software by Ansoft.
Through-silicon-via （TSV） to TSV crosstalk noise is one of the key factors affecting the signal integrity of three- dimensional integrated circuits （3D ICs）. Based on the frequency dependent equivalent electrical parameters for the TSV channel, an analytical crosstalk noise model is established to capture the TSV induced crosstalk noise. The impact of various design parameters including insulation dielectric, via pitch, via height, silicon conductivity, and terminal impedance on the crosstalk noise is analyzed with the proposed model. Two approaches are proposed to alleviate the TSV noise, namely, driver sizing and via shielding, and the SPICE results show 241 rnV and 379 mV reductions in the peak noise voltage, respectively.
In high speed three-dimensional integrated circuits （3D ICs）, through silicon via （TSV） insertion causes impedance discontinuities along the interconnect-TSV channel that results in signal reflection. As demonstrated for a two-plane interconnect structure connected by a TSV, we incorporate an appropriate capacitance at the junction to mitigate the signal reflection with gigascale frequencies. Based on 65 nm technology and S-parameter analysis, the decrease of signal reflection can be 189% at the tuned frequency of 5 GHz. Extending this method to the five-plane interconnect structure further, the reduction of signal reflection can achieve 400%. So we could broaden this method to any multilevel 3D interconnect structures. This method can also be applied to a circuit with tunable operating frequencies by digitally connecting the corresponding matching capacitance into the circuit through switches. There are remarkable improvements of the quality of the transmitting signals.
Metallic carbon nanotubes （CNTs） have been proposed as a promising alternative to Cu interconnects in future integrated circuits （ICs） for their remarkable conductive, mechanical and thermal properties. Compact equiv alent circuit models for single-walled carbon nanotube （SWCNT） bundles are described, and the performance of SWCNT bundle interconnects is evaluated and compared with traditional Cu interconnects at different interconnect levels for through-silicon-via-based three dimensional （3D） ICs. It is shown that at a local level, CNT interconnects exhibit lower signal delay and smaller optimal wire size. At intermediate and global levels, the delay improvement becomes more significant with technology scaling and increasing wire lengths. For 1 mm intermediate and 10 mm global level interconnects, the delay of SWCNT bundles is only 49.49% and 52.82% that of the Cu wires, respec tively.
