Contact metals?

By: Joe Hodge (4jhodge.delete@this.gmail.com), April 17, 2017 6:48 pm
Room: Moderated Discussions
tarlinian (tarlinian.delete@this.gmail.com) on April 14, 2017 8:59 am wrote:
> David Kanter (dkanter.delete@this.realworldtech.com) on April 14, 2017 8:26 am wrote:
> > David Kanter (dkanter.delete@this.realworldtech.com) on April 14, 2017 7:03 am wrote:
> > > DeMonet (abc.delete@this.def.ghi) on April 13, 2017 6:59 pm wrote:
> > > > I was just reading through the single dummy gate thread and realized
> > > > I had several questions about contact over active gate.
> > > >
> > > > Is increased density likely the primary reason for COAG? Or would it likely have other
> > > > benefits? I remember a slide from a couple years ago presented by Bill Holt (and later
> > > > referenced on this forum) that showed that the next major development after finfets would
> > > > be "Workfunction based Vt tuning" or something like that. Could COAG be related?
> > > >
> > > > What about lower contact resistance?
> > >
> > > I suspect it's a combination of both. Contact resistance is definitely a performance challenge
> > > these days. COAG may also be a prerequisite to doing new transistor architectures such as GAA.
> > >
> > > > Are there different alternative implementations of COAG that have been discussed in
> > > > the literature and Intel simply chose among them? Or is it something wholly new?
> > > >
> > > > Based on my somewhat brief survey COAG seemed to have come out of left field -- there are a couple of Intel
> > > > patents from several years ago that seem to refer to it, but I haven't seen it discussed much elsewhere.
> > >
> > > I haven't seen it in the literature much myself. Maybe Tarlinian has some thoughts?
> > >
> > > David
> >
> > Also, FWIW, Intel definitely did not want to disclose the materials used for the
> > contact at all. They were pretty sensitive about it. I suspect they also made some
> > changes to the interconnect stack, but did not want to disclose those either.
> >
> > David
>
> I don't think contact over active gate is anything specifically related to performance. (Workfunction
> metal based Vt tuning is completely unrelated and has to do with the metals used in the gate...you can
> add metals with different workfunctions to the gate stack to adjust Vt. There may be performance benefits
> from contact over active gate that result from a shorter conductive path. I think if you are looking
> for "alternative" reasons to implement CoAG, it's likely that whatever they are doing to implement involves
> some self-aligned process that also reduces litho requirements for equivalent density contact/LI layers.
> The S/D contacts were self aligned since 22 nm. Self-aligning the gate contacts is the next logical step.
> I expect you will see a basic process scheme published once chips are released.

My brain hurts trying to follow this, but would a shorter conductive path increase leakage current at a constant voltage?
< Previous Post in ThreadNext Post in Thread >
TopicPosted ByDate
Contact Over Active GateDeMonet2017/04/13 06:59 PM
  Contact Over Active GateDavid Kanter2017/04/14 07:03 AM
    Contact metals?David Kanter2017/04/14 08:26 AM
      Contact metals?tarlinian2017/04/14 08:59 AM
        Contact metals?DeMonet2017/04/14 02:42 PM
        Contact metals?Joe Hodge2017/04/17 06:48 PM
    Contact Over Active Gatetarlinian2017/04/14 09:42 AM
Reply to this Topic
Name:
Email:
Topic:
Body: No Text
How do you spell green?