[labnetwork] Copper Through-Glass Vias on fused silica

[Christopher Alpha]

Hi Folks,


(RE: Mac's comments)


We often do Pt seed layers by ALD for via Cu electroplating fills...you 
can get voids though depending on aspect ratio and how you drive the 
plating.  If a a hollow wire is OK this is fine to do in this fashion.


If you want a full solid fill we have done Au or Pt sputter on only one 
side of a very high aspect ratio via and basically try to sputter the 
hole mostly if not fully closed and use a plating fixture that makes 
contact on the back side with the Pt seed, leaving the other side 
exposed in the plating bath, then you can get a nice fill by plating 
from the bottom of the via up.  Overtop the via/wafer's surface, and CMP 
everything clean and flat (remove seed either selectively with wet 
chemistry if possible or CMP that off too).


Your mileage may vary...electroplating is a dark art.


Best,

C

-- 
Christopher Alpha
User Program Manager
Cornell Nanoscale Facility
Cornell University
250 Duffield Hall
343 Campus Rd
Ithaca, NY 14853
tel (607)254-4913

Any email from me will always really be from me without any AI "assistance"


On 2/26/25 13:04, Hathaway, Malcolm R wrote:
> Hi Larry,
>
> As it happens, Cu is not an established ALD process yet (that I am 
> aware of).
>
> Pt is common, but for your purposes (filling a 20um diameter hole) Pt 
> would likely too expensive.  In fact, ALD is probably not a viable 
> option because you will need a layer slightly thicker than the radius 
> of the hole (i.e. 10 um) and ALD of almost anything in a 10um 
> thickness will take 2-3 weeks (at least in our systems).  We tend to 
> speak of dep thickness/cycle (~1A) because to speak of rate in 
> thickness/min is too embarrassing!
>
> Which is to say, ALD is slow.
>
> Typical films are under 200nm.
>
> Material-wise, Pt, Ru, and some other conductors are available, but 
> they tend to be on the (even) slower side.  Some newer systems will 
> get cycle times down to achieve 20-30 cy/min, but still nothing 
> approaching CVD growth rates.
>
> A Pt liner (or other conductive material) with Cu electroplating is 
> much closer to a standard process for this sort of thing. And if you 
> wanted to get fancy you could even consider selective W CVD, but that 
> would require a bit of process development.
>
> I'm happy to chat about this off-line if you like.  Hopefully others 
> with proper electroplating experience will chime in shortly.
>
>
> Mac
> Harvard CNS
>
>
> ------------------------------------------------------------------------
> *From:* labnetwork <labnetwork-bounces at mtl.mit.edu> on behalf of Hess, 
> Larry A. (GSFC-5530) <larry.hess-1 at nasa.gov>
> *Sent:* Wednesday, February 26, 2025 9:36 AM
> *To:* labnetwork at mtl.mit.edu <labnetwork at mtl.mit.edu>
> *Subject:* [labnetwork] Copper Through-Glass Vias on fused silica
>
> Hi,
>
> I am looking for a (hopefully established) source to deposit ALD 
> copper on Though-Glass Vias.
>
> The vias are about 20um in diameter with an aspect ratio of about 10:1.
>
> I would also consider Cu electroplating( after an appropriate ALD 
> adhesion/base layer), but I’m not sure how feasible that is given the 
> via diameter and aspect ratio.
>
> Thanks for your help!
>
> Larry
>
> Larry A. Hess, PhD
>
> NASA
>
> Goddard Space Flight Center
>
> 8800 Greenbelt Road
>
> Code 553, Detector Systems Branch
>
> Building 11, Room E011
>
> Greenbelt, MD 20771
>
> 301-286-0259 (P)
>
> 301-286-1672 (F)
>
>
> _______________________________________________
> labnetwork mailing list
> labnetwork at mtl.mit.edu
> https://mtl.mit.edu/mailman/listinfo.cgi/labnetwork
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mtl.mit.edu/pipermail/labnetwork/attachments/20250227/7db43649/attachment.html>