[labnetwork] Flooded ICP-RIE

[Roy Cork]

Hi Peter,

Wow what a mess. I guess if the fluid is Hex ID it’s probably an Oxford ICP, what are the bakeout possibilities with that? On our SPTS ICP’s we can heat the sidewalls (approx. 70deg) and lid (50deg) and of course the electrode but that’s it (not enough I guess to evaporate the coolant). Down in the lower bellows (slice lift and electrode area) is not that hot and is away from the plasma, it is always the place where we get most polymer dep so I guess that indicates it’s the coldest part of the chamber.

Reading your post took me back a few years when I worked as an SPTS service engineer. I was working on an oxide etcher (AOE) screwing an electro static chuck to a copper electrode via a SS clamping ring. The screws sent were 12mm long instead of 10mm and where long enough to perforate the cooling cavity. When the chamber was closed Sytherm XLT oil was able to enter the chamber via the hole in the vented screws and drip down to the bottom section (the bellows were particularly bad). Of course we weren’t able to pump the chamber down and it wasn’t obvious when we opened the chamber as there wasn’t a “pool” of oil “just” a coating (the leak was a slow drip). We were able to recover the system by doing the following:

·         Manually strip the tool down and clean all parts in isopropanol and ultra-sonic bath. Take care with the bellows if you do this as they are fragile and easily damaged. We were able to save the MAG 2000 turbo pump by removing it from the system completely and poured 2-3L of IPA directly through it. We then dried it as best we could and re-mounted it.

·         Re-assemble the system and connect an RGA to the chamber. If you can try and obtain a reference RGA scan of a clean chamber

·         Pump the system down manually and heat it up. Here we were lucky, because it was an oxide etcher we were able to reach approx. 150 degs.

·         Leave it – monitor the RGA scan (we could see the higher peaks which were characteristic of the oil)

·         Leave it – I think we pumped for around a week whilst it was hot)

·         Eventually we were able to say the leak up rate and base pressure were good and the system became operational again
If the manufacturer has created this problem during a paid service visit I think you should insist they perform the strip down and clean, RGA scans and also provide some kind of warranty on future failure.

With regards the scrubber if it’s a CS Clean systems granulate system just close the inlet outlet valves, disconnect it and return it to FAB Detect maybe?? They can deal with it safely there. Be careful with the inlet and out pipes and where full breathing apparatus. If they are flexible bellows from the pump I would throw them out and replace them, be careful there is no liquid in the pipe when you remove it.

Good luck,
Best regards,

Roy Cork
DTU Danchip

From: labnetwork-bounces at mtl.mit.edu [mailto:labnetwork-bounces at mtl.mit.edu] On Behalf Of Paolini, Steven
Sent: 22. august 2016 22:53
To: Peter Köllensperger; labnetwork at mtl.mit.edu
Subject: Re: [labnetwork] Flooded ICP-RIE

Peter,
  Sorry for the bad luck.
I think your loadlock pump hydraulically locked. in other words, it swallowed enough water that when it went to compress the inlet “gas”, there was no compression because of the fluid. To be honest, it’s probable trashed unless it’s a rotary vane pump.
I tend to agree with Tom Britton and Rich Morrison’s responses in that you have a major clean up on your hands and it must be done meticulously. I would give the inside of the lift pin bellows special attention since the folds can hold a lot of contaminants. It’s probably best to remove all components that have been wetted and solvent clean them before re-assembly.
Best of luck,
  Steve Paolini
Equipment Dood
Harvard University Center for Nanoscale Systems.

From: labnetwork-bounces at mtl.mit.edu<mailto:labnetwork-bounces at mtl.mit.edu> [mailto:labnetwork-bounces at mtl.mit.edu] On Behalf Of Peter Köllensperger
Sent: Monday, August 22, 2016 4:15 AM
To: labnetwork at mtl.mit.edu<mailto:labnetwork at mtl.mit.edu>
Subject: [labnetwork] Flooded ICP-RIE

Dear Labnetwork Members,

I’ve just recently started as process group leader here at NTNU NanoLab in Trondheim, Norway, but have been following the discussions on the mailing list with interest.

I now have a question of my own and would very much appreciate any thoughts or advice you may be able to share.

We are currently having preventive maintenance of our ICP-RIE systems carried out by the manufacturer.
Unfortunately one of their engineers mixed up the helium backing line with the coolant line from the chiller unit. When the chiller was switched on, it pumped ca. 1-2 litres of a water and Hexid heat transfer fluid into the main chamber and loadlock. The system was not under vacuum at the time. The RIE section was completely filled with water and the main chamber also had water in it.  The lifting pins assembly including bellows, the CM gauge the pirani gauge as well as the He line were completely submerged.
The engineers realised what was going on after the chiller gave a low level alarm and they saw the coolant level continue to drop. They stopped the chiller, opened the chamber, and mopped up the liquid with PIG pads and cleanroom cloths where accessible. They then switched on both the roughing and loadlock pump.
When the loadlock pump was switched on it died immediately. They realised that water vapour shouldn’t enter the scrubber and removed the exhaust line from the scrubber, but it is likely that some water vapour ++ entered the scrubber as steam was coming out of the exhaust. This means the entire system has been exposed to water/coolant mix, possibly with exception of the turbo pump, although there may have been some backstreaming from the main roughing pump.

Pumping the chamber from Thursday<x-apple-data-detectors://1> to the following Tuesday improved the vacuum from 1x10e-2 Torr with a leakup rate of 20 mTorr/min to 5x10e-5 Torr with a leakup rate of 5 mTorr/min. Specification is less than 1 mTorr/min.

Needless to say this causes a number of issues, some of which we may not even have thought to consider.

In our view, the issues can be divided into immediate Health and Safety concerns, and more long term concerns regarding the performance of an ICP RIE unit that has been flooded


1.       Regarding H&S issues we’re concerned about are the mixing of SF6, CHF3, Cl2 and BCl3 with water in the scrubber.

We will send the scrubber for a refill and check. We have contacted the vendor of the scrubber, his main concern is clogging of the granulate. Is there anything else we should pay attention to?


2.       Can a chamber be sufficiently reconditioned by in situ by cleaning, baking and running plasmas or will the coolant inside necessitate a more complete disassembly and clean?
Has anyone had a similar problem and how did you proceed? How would you suggest we ask the manufacturer to verify that there are no long-term detrimental effects on the instrument, the pumps or the processes run on the instrument?


I'd appreciate any thoughts you may have surrounding this issue.

Kind regards,

Peter



––––––––––––––––––––––––––––––––––––––––––––––

Dr. Peter A. Köllensperger
NTNU NanoLab
Sem Sælandsvei 14
Norwegian University of Science and Technology (NTNU)
7491 Trondheim







-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mtl.mit.edu/pipermail/labnetwork/attachments/20160823/e3ab41d1/attachment.html>