Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/mate60-pro-finfet-tem.20636/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Mate60 pro Finfet TEM

jensy

New member
This is Mate60 pro TEM. I have one question, what is the protrusion at the circled area? In my opinion, it should not exist in finfet devices, but why haven't they been removed?
Thank you.
 

Attachments

  • 14D69E0C-74D8-4BA0-8FB9-B2AA2CF7A9C2.jpeg
    14D69E0C-74D8-4BA0-8FB9-B2AA2CF7A9C2.jpeg
    39.4 KB · Views: 282
These protrusions appear to only exist between Pch and Nch. (The color of the epitaxial layer appears different for Nch and Pch.)
Therefore, these protrusions are probably remnants of the overlapping areas of lithography when making the epitaxial layers of Nch and Pch separately. When etching the epitaxial layer mask, the STI SiO2 is over-etched and dented, but the overlapping areas of the Pch and Nch lithography are not etched.
 
These protrusions appear to only exist between Pch and Nch. (The color of the epitaxial layer appears different for Nch and Pch.)
Therefore, these protrusions are probably remnants of the overlapping areas of lithography when making the epitaxial layers of Nch and Pch separately. When etching the epitaxial layer mask, the STI SiO2 is over-etched and dented, but the overlapping areas of the Pch and Nch lithography are not etched.
Thank you. Can I understand that after the formation of n or p-type epitaxy, when one of the epitaxial layers is removed using a wet cleaning process, because of a low selectivity over STI SiO2, resulting in over etching of non lithography overlapping area.
Will this residual morphology have any negative impact on subsequent process integration or devices?
 
Thank you. Can I understand that after the formation of n or p-type epitaxy, when one of the epitaxial layers is removed using a wet cleaning process, because of a low selectivity over STI SiO2, resulting in over etching of non lithography overlapping area.
Will this residual morphology have any negative impact on subsequent process integration or devices?

When finfets were only starting, and geometry was bigger, there was an issue of fin bending depending on oxide thickness, and fin height. The early trick was to leave a bit of oxide during STI recess so it pushes back in the direction opposite to bending.

I have not seen that being used in production finfets after.
 
Thank you. Can I understand that after the formation of n or p-type epitaxy, when one of the epitaxial layers is removed using a wet cleaning process, because of a low selectivity over STI SiO2, resulting in over etching of non lithography overlapping area.
Will this residual morphology have any negative impact on subsequent process integration or devices?
I don't think that they etch away unnecessary epi regions after the epi process. The protrusion has nothing to do with the fin. I think that before (not after) epi process, a kind of SiO2 or SiN mask layer is applied to prevent epi growth in unnecessary areas. Unnecessary area means, for example, Nch region for the Pch epitaxial process. I mean the N/P overlap area of that mask opening litho is the cause of protrusion.
I do not know the subsequent effects of this protrusion.
 
Last edited:
I agree with your viewpoint, but I think there is another possible method, which is to first perform epitaxial growth on the entire wafer, and then use mask etching to remove unnecessary regions, such as retaining the p-type and removing the n-type. In theory, this approach should be feasible, but when removing the mask material of the n-type region, there may be issues with low selectivity from etching and wet cleaning over underlying layers, leading to over recess of the STI and Fin surfaces,thus STI protrusions. Then another n-type epitaxial material may be made in the n-type region. Comparing the two options, the second one may be more difficult to control.
 
I agree with your viewpoint, but I think there is another possible method, which is to first perform epitaxial growth on the entire wafer, and then use mask etching to remove unnecessary regions, such as retaining the p-type and removing the n-type. In theory, this approach should be feasible, but when removing the mask material of the n-type region, there may be issues with low selectivity from etching and wet cleaning over underlying layers, leading to over recess of the STI and Fin surfaces,thus STI protrusions. Then another n-type epitaxial material may be made in the n-type region. Comparing the two options, the second one may be more difficult to control.
I agree that various process integrations are possible around epitaxial layers. And I think that the approach you mentioned (starting from a blank epitaxial layer) is also interesting.
 
I agree that various process integrations are possible around epitaxial layers. And I think that the approach you mentioned (starting from a blank epitaxial layer) is also interesting.
I believe later processes somehow managed to remove the need for multi-stage STI recess, but SMIC may still be using that
 
Last edited:
Back
Top