3D structure and process

The Staircase Contact

10 sourced facts on The Staircase Contact, each tied to the source it comes from.

10 facts filed under 3D structure and process: all 10 are reported by a second source. Each was checked against the source’s own words, so its figures appear there, and each names where it comes from. The idea is taught in Part 4, 3D NAND Stacking.

Secondary sources

  • Trim-and-etch staircase: one mask can in theory make many steps, resist trim limits it, and staircase area grows in proportion to layer count.

    Published result3D Flash Memories

  • MiLC binary-sum staircase for vertical-gate 3D NAND: 8 layers need 3 masks instead of 8, and doubling the layers adds one mask.

    Published resultFigures: 8 layers, 3 masks3D Flash Memories

  • Routing 16 layers within the pitch of two pillar rows is a different job than across eight rows.

    Published resultFigures: 16 layers / 2 rows = 8 wires; / 8 rows = 23D Flash Memories

  • Program and read disturb spread along the plate in proportion to pillars per control-gate plate; slits trade density against disturb and block size.

    Published result3D Flash Memories

  • Trim-and-etch staircase: one lithographic mask could in theory make all the steps, but control gets difficult as layers grow, and the staircase area grows linearly with layer count.

    Published result3D Flash Memories

  • MiLC staircase for vertical-gate 3D NAND: for 8 layers, a superimposition of three masks opens each polysilicon layer.

    Published resultFigures: 8 layers, 3 masks3D Flash Memories

  • With the MiLC staircase, doubling the number of layers and contacts takes one extra mask.

    Published resultFigures: 1 mask per doubling3D Flash Memories

  • Connecting 16 layers within the pitch of 2 rows of pillars is a different job from connecting them within the pitch of 8 rows.

    Published resultFigures: 16 layers; 2 and 8 rows3D Flash Memories

  • Disturb is proportional to the number of pillars per control-gate layer, so slits that cut the layer into smaller blocks limit the spread of program and read disturb.

    Published result3D Flash Memories

  • The downside of array slits is their cost in silicon area, because the matrix becomes wider along the Y direction.

    Published result3D Flash Memories

Sources

Primary is the original paper, standard or datasheet. Secondary is a review, a vendor summary or a news report. Measured is my own test data.

Show all 1 sources
  • Secondary3D Flash MemoriesSpringer, 2016