[labnetwork] Contamination control in a UHV co-sputtering system (AJA Orion)

[Gerhard Ulbricht]

Hello Filippo.


All sputter systems are different but here are my experiences:


I never had any issues with real cross-sputtering (material from a 
neighboring, non active target being accidentally deposited). Typical 
sputter targets have grounded shields / chimneys around them and a 
shutter on top, together with the plasma forming magnets underneath your 
active target this reliably prevents non-active targets from sputtering.
There are two risks though: If you have several densely package sputter 
guns, you risk depositing some of material A during sputtering onto 
inactive gun B. If you then sputter from gun B later you will get 
material B + a bit of A. This is pressure dependent but especially at 
higher Ar pressures some of your sputtered atoms will make a U-turn. To 
prevent this make sure the gap between shutter and chimney of all guns 
is as small as possible, and always run your target a bit against the 
closed shutter for cleaning.
A similar issue can arise if you have two or more guns active 
simultaneously: some material from A can deposit on B and will thus be 
co-sputtered by gun B. You will not even notice that during your process 
as you want to deposit both materials anyways. But after co-depositing 
you should run all targets for some minutes of sputter time in an empty 
chamber to make sure their targets are clean again.


Lithographic resist in a UHV sputter system: Depends on your definition 
of UHV. If for you UHV means 10^-10 mbar or better, you need to bake out 
anything you bring into the chamber, including your resist covered 
samples. That will likely destroy / reflow /... the resist anyways. If 
you are fine with 10^-8 mbar and therefore do not require strict 
bake-outs, most lithographic resists I have used are totally fine to be 
sputtered upon (there are many resist though, so there could be 
exceptions). I am referring to resists that have seen post-exposure 
bake-out before going into the vacuum chamber, reducing the risk of 
out-gassing during sputtering.
But if very sensitive samples are run in a sputter system, resist is 
usually banned, not necessarily because of the resist itself but because 
a resist covered sample by definition is not as clean as possible (it 
can't be sonicated / plasma-cleaned / ...) and therefore risks 
contaminating the chamber with anything that accidentally sits on the 
resist.


The use of Kapton tape in a 10^-8 mbar or better sputter system is not 
recommended: The sample surface will likely get warm during sputtering, 
risking the glue underneath the Kapton to evaporate and deposit anywhere 
in the system, potentially contaminating future samples. I have never 
tried Crystalbond myself though.


Hope that helps.
Gerhard.




On 11/08/2026 03:13, #MARTINELLI FILIPPO# wrote:
> Dear labnetwork community,
>
> I have recently been put in charge of a UHV magnetron sputtering 
> system (AJA Orion) that I need to maintain and operate. The machine is 
> dedicated to co-sputtering of two metals in a nitrogen environment.
>
> The previous operator strongly advised me to be careful about what 
> goes into the chamber, as contamination can affect both the chamber 
> condition and the properties of the deposited material. However, she 
> was not very specific about what exactly causes contamination, and I 
> am unsure how restrictive I should be with the materials I load.
>
> I have a few questions I would appreciate your opinion on:
>
> 1. I read that cross-sputtering can occur between nearby substrates, 
> or between the substrate and the chamber. In your opinion, are there 
> materials that are particularly prone to cross-sputtering and hard to 
> remove with standard cleaning routines, and that should therefore be 
> avoided in the chamber altogether?
>
> 2. Following on from that, what is your take on lithography resists?
>
> 3. How about mounting materials such as Kapton tape or Crystalbond?
>
> I am quite new to this, so any suggestion or recommendation is 
> welcome. I look forward to hearing your opinions.
>
> Cheers,
> Filippo
>
> _______________________________________________
> 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/20260811/bed2b73b/attachment.html>