Skip to content

THE COURSE / M8 INTEGRATION / M8.2

The stack and the fin

After this lesson you can:

  • Say what each layer of the silicon and silicon-germanium stack becomes.
  • Explain why the silicon beneath the stack is doped first, and why the stack is grown cool.
  • Work out how much channel width a stack of sheets gives, and say what the fat-fin problem is.

Loading the lesson…

Sources

  1. Q. Zhang, Y. Zhang, Y. Luo and H. Yin, "New structure transistors for advanced technology node CMOS ICs," National Science Review 11, nwae008 (2024). Stacked nanosheet transistors: about 80% of the flow shared with a gate-last FinFET, plus four new steps (sub-fin doping, Si/SiGe superlattice epitaxy, inner spacers, channel release); the superlattice grown below 650 °C; inner spacers of more than ten steps; SiGe removed at over 100:1; the sub-fin's leakage and bottom dielectric isolation; wider sheets for more current; the gate between the sheets. FinFETs used from the 22 nm node to 3 nm; fins grown thinner and taller bring reduced mobility, random doping fluctuation and Vt differences, and around 3 nm short-channel effects. GAA nanosheets: better electrostatics than FinFETs, more effective width from stacked sheets, and a broad allowable width for flexible design; Samsung's stacked three-sheet GAAFETs with 12 nm gates; an imec 2016 nanowire with a 24 nm gate reached 65 mV/decade and DIBL of 42 mV/V. Channel doping much reduced in FinFETs and GAAFETs. CFET stacks the n- and p-type transistors vertically, removing the n-to-p spacing, toward cells of four or even three tracks.