[SOLVED] CONT is VIA for Poly to M1 and also for Nwell?

[gujaratibhai]

Hi folks,

CONT (contact) is a via for M1 to Poly connections. And CONT is also the via to connect to Deep-Nwell or PIMP layers. There exists a SiO2 layer exist underneath the poly, correct? Thus CONT for Poly must be shorter than the CONT to connect to the a Deep NWELL underneath? When I look at of CMOS cross sections, it's hard to distinguish.

In the layout software, CONT is CONT but I always wonder that the height for CONT for poly must be different than the one used to connect to the substrate. If you can explain or point me to a cross sectional view somewhere on the web specifically addressing CONT layer, I'd greatly appreciate it.

Thanks,

 

[rangermad]

Generically speaking, poly stops the etching of the hole for your contact. Other contacts are etched at the same time and they will go all they way down to silicon. It is time based.

There is no direct connection from contact to dnwell. The connections go metal, contact, N+, nwell, dnwell.

Generically spewaking, of course.

 

[erikl]

... CONT for Poly must be shorter than the CONT to connect to the a Deep NWELL underneath?

You're right, CONT to poly is actually a bit shorter than the CONT to the silicon level, s. this symbolic cross section:


and e.g. here for a 0.18µm process:


This small difference doesn't matter at all, because metal fillings can be made much deeper, s.
this deep trench metal filling, about 20µm deep.

 

[gujaratibhai]

Erikl,

Thanks.

Whenever I see "erikl" responded, I generally read the response just because.

Thank you for being wealth of information and actually offering your help to many!

Regards,
Gujaratibhai