[labnetwork] LN2 vs N2 generation on site

Fri Mar 27 14:08:13 EDT 2015

Hi Everyone,

I would like to share with you some details. We have 2 3000 gallon tanks. Tank one feeds the Microfabrication Laboratory and is filled every other day. The 2nd tank feeds 3 different labs, Assembly, MCM and 5th Floor and is filled every 4th day, this systems is all SS from the tank to point if use. We do not use purifiers so I expect that even though the LN2 is 99.999% I have something less at the point of use.

The Microfab is mostly dry boxes, N2 guns, purge for pumps, purge for 3 furnace tubes (oxidation) and 3 LPCVD system, poly, nitride and SiC. We run all dry pumps and have not had any problems using this gas stream. We shut of the SRD at night and weekend to save gas but I am still at a high flow. From the tank to the Microfab loop is all SS, the Microfab loop is  oxygen free EP Copper silver soldered.

Microfab runs 23 cuft to 34cuft/min  (SRD dependant)
MCM runs 3-6 cuft/min  depending on tools usage
Assembly 1-16 cuft/min depending on tool usage
5th Floor 1-5 cuft/min depending on tool usage

The MCM, Assembly and 5th floor need oxygen free Nitrogen with no moisture;  so my guess here 99.99% Nitrogen or 99.9%?
The Microfabrication should be 99.99% or does it need 99.999%. We have never had our existing system analyzed, any ideas here?

[cid:image002.png at 01D06897.7677FDB0]

[cid:image004.png at 01D06897.7677FDB0]

Thanks for your comments so far and I looked forward to more discussion. BTW after fees and surcharges I pay $1.23 per 100cuft and I use 1.45million cuft per month (data from flow meters). The company delivers the equivalent of 2.6million cuft so I lose to evaporation and cooling 1million cuft per month, does this sound right or do I have a big leak some place?

Rick

From: John Shott [mailto:shott at stanford.edu]
Sent: Friday, March 27, 2015 11:55 AM
To: Tribble, Thomas; Morrison, Richard H., Jr.; labnetwork at mtl.mit.edu; Nibarger, John
Subject: Re: [labnetwork] LN2 vs N2 generation on site

Rick:

This is an excellent question and I thank Tom Tribble and John Nibarger for their initial contributions to this topic ... and others that will undoubtedly soon follow.

First, let me get my remedial class question out of the way:  Tom, you've got me feeling like a rank beginner.  I must confess to being unfamiliar with the 1δ, 3δ, and 6δ notation for gas purity.  Is this the equivalent of what I know as dot notation? In other words, so they correspond to 1-, 3-, and 6-nines purity that is 1.0, 3.0, and 6.0 in dot notation?  Where 99.5% purity is 2.5?  Or is δ notation something different?

I would like to add confirmation of your "unofficial audit" that found that 10% of your usage demands high purity.
We also have a 9000 gallon LN2 tank (which, sadly, gets filled every fourth day).  We have two sets of vaporizers on it: the big set supplies the copper-piped "house N2" and the smaller vaporizers that also go through a switchable set of purifiers and then feed our stainless-piped "UHP N2" system.  Nouse N2 usage is dominated by pump purge.  We have totalizing flow meters on each main line and typically observe that the UHP N2 consumption is 130-200 SLM and the house N2 consumption is 2500-2800 SLM.  In other words our total nitrogen flow is very close to 3000 SLM (or about 4.5M - 5.0M SCF per month!). In other words, our high-purity consumption is in the range of 5-7% of our total consumption.  Note: we actually use the house nitrogen for our SRDs ... if we moved that intermittent high-flow to UHP N2, it probably gets us close to, but no higher than, 10%.

My question to the labnetwork community is what level of nitrogen purity would you want for pump purges of some of our nastier processes?  For example, things like DCS and ammonia for LPCVD nitride?  Hydrogen, germane, silane, plus dopants for an epitaxial process?  is 1.0 nitrogen good enough?  The way that I look at it, that is only a 50% reduction of oxygen as compared to what you find in air.  How about 2.0 nitrogen (99%), good enough?

And one question to John Nibarger:  is the moisture content of your generated nitrogen controlled by the moisture content of your feed CDA?  In other words, with your incoming CDA dew point of -40C is the dew point of your generated nitrogen also approximately -40C?  If the incoming CDA were drier (say -55C to -60C) do you expect that your generated nitrogen would also be that much drier?  I ask because for pumps running corrosives such as HCl, Cl2, BCl3, and HBr it would seem that dryness of the purge gas would be more important than the background oxygen content.

Thank you all,

John

On 3/24/2015 1:31 PM, Tribble, Thomas wrote:
Richard,

We haven’t switched to onsite generation for the fab, but we (Harvard) have looked at it.  The Harvard fab has a 9000 gallon LN2- 3δ storage that gets re-filled approximately once a week.  The storage facility occupies approximately 300 GSF, and there  is no more real estate available for this purpose.  Based upon a much smaller on-site generator recently completed, on-site generation to replace the fab storage facility would require more real estate than is available.  Part of the demand for space is the self-imposed requirement for N+1 redundancy.  This also affects the capital budget.

But if saving money is your focus, I suggest that your discussion should be focused on the purity of your gas requirements. Locally generated 1δ gas is inexpensive cheap to produce. Locally generated 3δ gas is moderately expensive to produce.  Locally generated 6δ gas is very expensive to produce.  From the refinery, 3δ gas is all that is produced and transported.  What we at Harvard use is 6δ gas (3δ gas run through a purifier).  An unofficial audit has shown that less than 10% of our usage actually demands 6δ purity.

Want to save money?  Figure out your demand by purity and segregate the flows (piping).  Provide your pump and gas cabinet purges with on-site generated 1δ gas.  Your facility will save.

Tom Tribble
From: labnetwork-bounces at mtl.mit.edu<mailto:labnetwork-bounces at mtl.mit.edu> [mailto:labnetwork-bounces at mtl.mit.edu] On Behalf Of Morrison, Richard H., Jr.
Sent: Tuesday, March 24, 2015 12:04 PM
To: labnetwork at mtl.mit.edu<mailto:labnetwork at mtl.mit.edu>
Subject: [labnetwork] LN2 vs N2 generation on site

Hi everyone,

I am paying >$220K per year for LN2 to generate Nitrogen for my fab. Have any of you guys switched to onsite Nitrogen generator?

Looking for ideas on:
Cost of system
Maintenance cost of system
 Quality of the Nitrogen.
Operating cost per CuFt of Nitrogen

Rick

Draper Laboratory
Principal  Member of the Technical Staff
Group Leader Microfabrication Operations
555 Technology Square
Cambridge Ma, 02139-3563

www.draper.com<http://www.draper.com>
rmorrison at draper.com<mailto:rmorrison at draper.com>
W 617-258-3420
C 508-930-3461

________________________________
Notice: This email and any attachments may contain proprietary (Draper non-public) and/or export-controlled information of Draper Laboratory. If you are not the intended recipient of this email, please immediately notify the sender by replying to this email and immediately destroy all copies of this email.
________________________________

_______________________________________________

labnetwork mailing list

labnetwork at mtl.mit.edu<mailto:labnetwork at mtl.mit.edu>

https://www-mtl.mit.edu/mailman/listinfo.cgi/labnetwork

________________________________
Notice: This email and any attachments may contain proprietary (Draper non-public) and/or export-controlled information of Draper Laboratory. If you are not the intended recipient of this email, please immediately notify the sender by replying to this email and immediately destroy all copies of this email.
________________________________
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mtl.mit.edu/pipermail/labnetwork/attachments/20150327/054f9a1d/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.png
Type: image/png
Size: 41506 bytes
Desc: image002.png
URL: <https://mtl.mit.edu/pipermail/labnetwork/attachments/20150327/054f9a1d/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image004.png
Type: image/png
Size: 38783 bytes
Desc: image004.png
URL: <https://mtl.mit.edu/pipermail/labnetwork/attachments/20150327/054f9a1d/attachment-0001.png>