https://open-source-silicon.dev logo
Join Slack
Powered by
So I was looking at zeptobar decap / deprroessing ...
# sky130
t
So I was looking at zeptobar decap / deprroessing of a TT4 chip :

https://s.zeptobars.com/mpw4-tt-ctrl-HD.jpg▾

and I'm wondering what we're seeing exactly. It seems to match the "diffusion" layer ( the polysilicon gates fell off apparently since they weren't held by anything after the HF). But I don't understand how. Because the diffusion layer in the GDS in practice should end up masked by the gates (i.e. implants are made after the gates are formed and so there is no diffusion under the gate, so it forms the channel), but here we see something that exactly matches the diffusion GDS layer and isn't masked by the gates, so what are we seeing ?
d
You can visually make out the diffusion area. Remember there's an etch of the oxide before the diffusion area is implanted with whatever ions are used. That etch mostly matches the diffusion area. After all the processing there is still a slightly visible boundary.
t
Yeah, some oxide is etched using the diffusion area mask, but after the implantation, the remainder is stripped isn't it ? I'm looking at https://web02.gonzaga.edu/faculty/talarico/EE406/20162017/Lectures/layout.pdf for instance (page 17/18/19).
d
Maybe so, I'm not sure. My info is a bit aged. I remember there being an indication of the diffusion. Not like a metal or anything solid, but enough to tell where it was. This was on SKY130, too.
Here's what one of the AI say....... 😀 1. Surface effects: In some cases, heavy doping can cause slight changes in the surface appearance of the wafer, such as a very subtle color change or surface texture difference. 2. Etch patterns: When the wafer is subjected to certain etching processes, the diffused areas may etch at different rates than the non-diffused areas, creating visible patterns. It's important to note that in most cases, the diffusion itself is not directly visible. Rather, it's the effects of diffusion or the results of specific analysis techniques that make the diffused regions detectable or "visible" in various ways.
t
Yeah, I can get why something would appear where there is actual diffusion. But what I don't get is why this isn't interrupted by the gates. Since whatever happens to the substrate because of the diffusion/implantation, this shouldn't happen under the gate (the whole self aligned thing).
I assume skywater doesn't publish the steps list for sky130 manufacturing ? 😅
👍 1
The only think I could see is Vt implants.
d
Could be.
l
Yes, that looks like the active region plus STI, since there's no more poly gates to define the transistors.
6 Views
Open in Slack
PreviousNext
Powered by