[labnetwork] HPM gas survey

[Weaver, John R]

Michael –
Following are our methods and my approach to these issues. They have been working very successfully for over 10 years now, and I believe in this approach. It is the culmination of 35 years in the semiconductor industry and service on the NFPA 318 Committee as a Principle Member prior to the development of this system for the Birck Nanotechnology Center at Purdue.
Please let me know if you need more details or the rationale for each of these decisions.
John

VMBs: We use VMBs throughout the facility, both for cleanroom and laboratories. VMBs are vented with detection in the vent line 24 inches above the VMB transition to ensure turbulence in the exhaust so that there is no “streaming” bypass of the detector. No lines are “teed” – everything is a separate feed from the VMB. I am using 4-stick VMBs. If I have three sticks in use, the fourth stick feeds a daisy-chained VMB, giving me three more sticks. I DO NOT use VMBs for 100% silane! Each feed line for 100% silane is a home run from the gas cabinet.

Coaxial Lines: ALL hazardous gases (3 or more on the old NFPA scale in any category) use coaxial piping. All delivery lines are welded and are of 316L SS tubing. I don’t use any Hastelloy for chlorinated gas lines, but do use Hastelloy trim for all valves in chlorinated gas lines. Secondary containment is 316L SS tubing, but I allow Swagelok connections in these containment lines. It lessens cost and facilitates repairs should you have a leak in the delivery line. I run a static mix of nitrogen and helium (for leak detection purposes) in the interstitial at half the delivery pressure of the gas. The interstitial pressure is monitored real-time. If the interstitial pressure increases, there is a leak in the delivery line. If the interstitial pressure decreases, there is a leak in the outer containment. In either case, a text message is sent to the appropriate people. If the interstitial pressure goes to zero, there has been a catastrophic failure and the facility is evacuated.

Hydrogen: I keep all my hydrogen cylinders outside the building. I had a fire marshal at a previous location who referred to a hydrogen incident as a “lid lifter.” That terminology made my blood run cold, and I am very conservative regarding nitrogen. I have found that having exterior cylinders, even in cold weather, is not too much of a burden. I use a truck-refillable microbulk for that purpose.

General: Our mantra is that all points where a hazardous gas is not in welded, doubly contained piping – e.g., gas cabinet, VMB, equipment – it is in an exhausted enclosure, and that enclosure is sniffed as mentioned above. No exceptions.

Heated Lines: We heat the panels and lines for all low vapor pressure gases.

Distribution Rooms: We have two concrete-block gas rooms, one for flammable and one for toxic gases. Flammable toxics go in the toxic room. All cylinders for hazardous gases are in individual cabinets with autopurge panels with individual purge cylinders. All cabinets are sprinklered. Pyrophoric gases are located in a bunker external to the building with remote purge panels. The bunker has three poured-concrete walls and a blow-out wall facing a fenced enclosure and a blow-out roof.

I hope this helps,
John

From: labnetwork-bounces at mtl.mit.edu [mailto:labnetwork-bounces at mtl.mit.edu] On Behalf Of Michael Khbeis
Sent: Tuesday, December 06, 2016 10:17 AM
To: labnetwork at mtl.mit.edu
Subject: [labnetwork] HPM gas survey

Dear Colleagues,

Once again, I am pitted against my capital projects group. This time regarding HPM gas infrastructure. Thanks to your prior input, we were able to obtain a proper toxic gas monitoring system. However, now I am in a code interpretation debate with their process piping consultants on the use of VMBs for multi-tool feeds and coaxial lines. They don't want either in our project. They specified hastelloy for Chlorine and BCl3 but single walled 316ss (after much debate that 304 was not sufficient) for all our other gasses. In ALL my prior sites we always used coaxial lines as a secondary containment for HPMs, but currently this project has no coax specified.

I am also in debate on whether or not 100% hydrogen (large 330cf cylinders) can be in the Cleanroom without a cabinet vs a bunker in a cabinet. Their interpretation is that an H-5 facility can store 9000cf; however, I believe they are ignoring the vented enclosure requirements.

We will have the following HPM gasses:
100% Silane
100% Hydrogen
Ammonia
Chlorine
Boron Trichloride
Phosphine
Diborane

Can you please weigh in on your sites use of:

1) coaxial lines, which gasses and if any HPMs are excluded or use single hastelloy

2) use of valve manifold boxes / vented enclosures for distributing HPMs to multiple tools (or anywhere else there's a mechanical connection)

3) location and enclosure or not for 100% hydrogen

And on a non-safety related note, we asked for panels with multi-stick feeds to each tool; however, the process piping designer just teed all of our lines from a single point and put a regulator at each tool instead to "save money". I am very concerned about LP gasses (e.g C4F8) being starved when multiple tools are running. We distinctly had issues with past lab with shared C4F8 bottles. Their recommendation is to ONLY heat the cylinder to increase vapor pressure, but given my limited knowledge, I believe with high flow rate and lacking heating of the panel and lines we will have gas condensation issues. To date, my input is still being dismissed so please state:

4) how you handle LP gas distribution to multiple tools

I thank you so much for taking the time out of your busy schedules to provide this input. This community/forum has been such a blessing.

Most gratefully,

Dr. Michael Khbeis
Washington Nanofab Facility
University of Washington
Fluke Hall, Box 352143
(O) 206.543.5101<tel:206.543.5101>
(C) 443.254.5192<tel:443.254.5192>
khbeis at uw.edu<mailto:khbeis at uw.edu>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mtl.mit.edu/pipermail/labnetwork/attachments/20161207/575a8c95/attachment.html>