[content] => 
    [params] => Array
            [0] => /forum/index.php?threads/lithographic-patterning-using-tilted-ion-implantation-tii.8663/

    [addOns] => Array
            [DL6/MLTP] => 13
            [Hampel/JobRunner] => 1030170
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000670
            [ThemeHouse/XPress] => 1010394
            [XF] => 2011072
            [XFI] => 1030270

    [wordpress] => /var/www/html

Lithographic Patterning using Tilted Ion Implantation (TII)

Tom Dillinger

To achieve an aggressive lithographic pitch in advanced process nodes, foundries have employed self-aligned double patterning (SADP), using sidewall spacers as the final masking layer. In the future, another iteration of this method will offer quad patterning, or SAQP.

The team at UC-Berkeley has been investigating an alternative approach to realizing an aggressive masking pitch, known as Tilted Ion Implantation (TII). As illustrated below, the approach uses "shadowing" of an Ar+ implant into an SiO2 layer, using a base pattern of a Hard Mask (HM) material as the implant block.

View attachment 18617

The implanted SiO2 areas will have a different (wet HF) etch rate than the unimplanted regions, allowing the final SiO2 masking material to be patterned between and undercut below the base HM material. The final SiO2 etch pattern and linewidth tolerances are a function of many parameters -- e.g., HM height and profile, implant angle, implant dose/energy, SiO2 layer thickness, etch selectivity between implanted and unimplanted regions.

A thorough description of the TII approach can be found in a recent IEEE Transactions on Electron Devices paper:

Sub-lithographic Patterning via Tilted Ion Implantation for Scaling Beyond the 7-nm Technology Node - IEEE Xplore Document

Last edited:

Li Yisuo

Awesome idea!
But the implant profile will leave a strong footing at bottom SiO2, which can be easily inherited by the hard mask.


How is this superior to using e-beams to pattern? With all the scale-up problems of e-beams?
I assume there's some scope for variation in that you have a different energy/wavelength trade-off and perhaps you can use lower energy ions, but scaling seems just as problematic as always.

After all, I assume the reasons why people are putting up with the endless pain of EUV rather than just chucking it in and going to e-beam still hold?

Fred Chen

ArF immersion has highest throughput but resolution can only be extended beyond ~40 nm by multiple patterning. It's already ~20 nm by double patterning. Ebeam and EUV throughput are too low; EUV is slower than ArF immersion double patterning (1500 =>750 (higher dose) wafers/day vs. 6000 wafer-passes/day); ebeam even slower than that (1-2 wafers/day or even less).
Last edited: