Air - Water Spectrum MS

[lusi]

Hey Pete...Will look at the forum archives a bit more...hopefully I find that post.

The column is a DB5, by the way.

Now I am seeing the Mz18 fluctuating about 60. And the MZ28 is fully at 100. So a definite drop in the Mz18.

This is so frustrating for me! I have no experience and the manual does not describe a situation like this.

[Peter Apps]

Hi Lusi

How rapid are the fluctuations in m/z 18 ? If they are over periods of several hours I suspect that changes in ambient humidity are the cause.

Just to check: you do have moisture and oxygen scrubbers on you carrrier gas, and they are not exhausted ?

Peter

[lusi]

Hi Peter,

The fluctuations were pretty rapid, as in while looking at the screen.
I did get to speak to the supplier about the overnight variation (rise in Mz28), he suggested maybe caused by condensation during the nights.
He also indicated that they did have a customer with a similar case; however the manufacturer indicated once the leak tests are good, the instruments can be used.

Our lines run out of the building into the room with the GC. We have an equatorial climate.

He also said it maybe a small leak. I changed the seal at the injector. So I am yet to run the overnight at 180. TO be honest I was afraid that there is a leak somewhere and it may ruin the cloumn if I raised the temp beyond 100..

The filter looks pretty good to me.

[Peter Apps]

Hi Peter,

The fluctuations were pretty rapid, as in while looking at the screen. OK, so this is not ambient humidity


I did get to speak to the supplier about the overnight variation (rise in Mz28), he suggested maybe caused by condensation during the nights.
He also indicated that they did have a customer with a similar case; however the manufacturer indicated once the leak tests are good, the instruments can be used. This is kind of obvious don't you think ? The question is whether you really have a leak, or is something else generating an m/z 28 signal that the software interprets as being a leak ?

Our lines run out of the building into the room with the GC. We have an equatorial climate. Does the sun shine on any of the gas lines (which I assume are made out of copper or stainless steel ?)

He also said it maybe a small leak. I changed the seal at the injector. So I am yet to run the overnight at 180. TO be honest I was afraid that there is a leak somewhere and it may ruin the cloumn if I raised the temp beyond 100..

The filter looks pretty good to me. How do you mean ?, does it have an indicator that changes colour ? What kind of "filter" is it

Have you checked all the GC plumbing with a leak seeker - NOT with the onboard leak diagnostics ?

Peter

[lusi]

Hi Peter,
The lines are copper and yes sun does get them to some extent.
The last few days have been pretty a mixture of overcast, sun and rains.
Yes we have an indicator filter. Green to grey with O2 and yellow to clear with moisture. It was the recommended (by supplier) one for the instrument.
Yes we checked the lines several times and no leaks.

[Peter Apps]

Hi Lusi

The possibilities are:

you have a real leak, and air is diffusing into the system. But you do not find a leak with a leak seeker - just confirm that you have checked all the lines going to the GC and all the plumbing inside the GC. Any leak that is too small to find with a leak seeker is too small to worry about.

You do not have a real leak. In which case you can use the MS.

There is no leak, but the gas in the cylinder has air and moisture in it, and the scrubber is not removing them. What purity grade of gas is it ?

Do you have any evidence at all, apart from what the software is telling you based on 18 and 28 that there really is a leak ?

Peter

[lusi]

The gas in ultra-high purity. Its about 99.99

I still cant find the leaks physically. I checked the lines and at the inlets. Is there a notorious place I am missing?

The automatic leak test on the GC (using the column attached to the injector) passed. The leak test of the autotune also passed.

I am tempted to change out the columns. Maybe the ferrule gave out? But then there would have had a vacuum leak?

[Peter Apps]

The gas in ultra-high purity. Its about 99.99 AHA ! this is not high purity by GC-MS standards. You need at least 99.999%, instrument grade, and then you still need to use a scrubber. If you cannot get that then you will have to put a high capacity oxygen and moisture scrubber just upstream of the GC, and change it frequently.

I still cant find the leaks physically. I checked the lines and at the inlets. Is there a notorious place I am missing? Does this include all the pipes and controllers inside the GC ? - just to be clear I mean the plumbing underneath the top cover and side panels

The automatic leak test on the GC (using the column attached to the injector) passed. The leak test of the autotune also passed. So why do you think that you have a leak ? - you will always see some 18 and 28

I am tempted to change out the columns. Maybe the ferrule gave out? But then there would have had a vacuum leak? Only gross leaks make a measurable difference to the vacuum in the MS, if the leak was that bad you would have burnt your filament by now

Peter

[lusi]

Sorry Peter, I did leave out an extra 9, so it is 99.999% But it has been sitting around for a couple of years. But the manufacturers indicated this is not a problem.

Lets say things take about 6months to reach to our labs - even longer. So we tend to order a few if we get the approval.

Well I did check the top once, but did not really take anything apart just remove the cover. Never did the side panels. Honestly never seen the suppliers remove the side panels.

Honestly I dont know if I have a leak. But somehow the system always develops a leak. Or what appears to be a leak. We use the graphs in the users manuals, so if the 28 is higher than the 18 then there is a possible leak.

Thanks for the explanation about the vacuum leak. What about the column leak test that is done on the GC? I always assume that once that passes the inlets and everything before the inlets are fine? Am I wrong in that assumption.

Sad thing training was limited and the tech support is a whole other issue...

[Peter Apps]

Sorry Peter, I did leave out an extra 9, so it is 99.999% But it has been sitting around for a couple of years. But the manufacturers indicated this is not a problem.

Lets say things take about 6months to reach to our labs - even longer I can certianly relate to that . So we tend to order a few if we get the approval.

Well I did check the top once, but did not really take anything apart just remove the cover. Never did the side panels. Honestly never seen the suppliers remove the side panels. taking the panels off is only necessary if they are blocking your access to the pipework and controllers so that you can leak check them. I do not have the particular model of GC that you use, but generically the gas connects somewhere at the back and there are a series of connections and controllers upstream and downstream of the inlet. You have to trace all of these frrm the point that the gas enters the GC at the back to the point where it leaves the GC via the split and septum purge vents and check every connection for leaks with the leak seeker. You should not dismantle any of the gas plumbing and DO NOT just tighten everything just in case it might be leaking

Honestly I dont know if I have a leak. But somehow the system always develops a leak. Or what appears to be a leak. We use the graphs in the users manuals, so if the 28 is higher than the 18 then there is a possible leak.As I have been trying to explain, there are other possible reasons for 28 to be higher than 18. If this is the only evidence of leak, then there is probably no leak

Thanks for the explanation about the vacuum leak. What about the column leak test that is done on the GC? I always assume that once that passes the inlets and everything before the inlets are fine? Am I wrong in that assumptionI do not know how the GC leak test works, presumably it is some kind of pressure holding test. I very much doubt that it can check anything upstream of the electronic gas controls i.e. any leak in the feed pipes and the internal plumbing of the GC between the connection at the back of the EFC, or anything at the downstream end of the column.

Sad thing training was limited and the tech support is a whole other issue...

Peter

[lusi]

Hi Peter,
Thanks for all the advice and feedback.
Yesterday I noticed that the graph is consistent with a leak.
Other than that I had no proof of a leak.
The technician was saying maybe the calibration gas is low.
Not sure how this would explain the Mz18 not being present.

[lusi]

So following was found (with a service engineer's help)
1) the ion source was not place properly inplace, in fact there is a screw there that is slack.
2) the line connecting the calibration compound to the transfer line need tightening a bit more
3) more calibration compound was added ( none was added in about 4yrs)

So the problem was solved by the time all of these were corrected.

[Peter Apps]

Hi Lusi

Thanks for the feedback.

Peter

[lusi]

No Problem Peter

Thanks for all your help.