GC-MS Baseline Issues

[EntoJoe]

Hello Chromatography Forum,

I'm a PhD student looking for a little bit of advice with regards to some issues that I have been experiencing with the GC-MS system I have been using for some of my analyses and was hoping someone here would be able to point me in the right direction as to solving them.

I am running an Agilent 7890B GC System with an Agilent 5977A MSD in order to analyse the volatile organic compounds emitted by a species of herbivorous mites. However, the machine was used by an inexperienced user without my knowledge and since I have returned to use it I am experiencing issues with my chromatograms - particularly peak size and baseline. Essentially the chromatograms and resulting spectra appear to be very weak, and the baseline is completely flat between peaks (i.e. there are no ions being detected). This is the case with both liquid and thermal desorption injections, however the liquid injections drop off to 0 after 4.00 mins while thermal desorption injects are always at 0. I have linked to a picture to demonstrate what I mean: http://imgur.com/FeRtajr. Prior analyses I have run on the system did not suffer from this lack of a baseline, and the peaks that are present are far smaller than I would expect.

Does anyone have any suggestions as to what might be causing this?

I'm sorry if this is a stupid question, but it has me stumped. If you require any further information or screenshots I will happily provide them.

Thanks in advance,

Joe

[Peter Apps]

There are several possibilities; drastic changes in flow or temperature on the GC would explain a rapid but not sharp downward baseline. Anything programmed into the MS such as switching from full scan to SIM would produce a sharp drop.

If this happens consistently then something in the operating programme has changed, so re-load your original method, or check through the current one and change it back to what it should be.

Peter

[EntoJoe]

Hi Peter,

Thanks for the response, it is much appreciated.

I have reviewed the methods and nothing appears to have changed as far as I can tell, yet I am still having issues.

Joe

[cleh]

Please post a spectrum of the elavated baseline.

[James_Ball]

What ions are present in the higher portion of the baseline at the beginning of the run?

If the major ions are m/z 44 and 40, then it could be that the nut on the column at the mass spec interface is loose. What will happen is that as the oven heats up the ferrule will expand and seal the connection, then when it cools it will begin to leak again.

Have you run an Autotune or Air/Water check since returning to using the instrument?

Also check to make sure there are no timed events that have been added that change the electron multiplier voltage or gain settings as the run progresses, these could cause similar problems as you have.

[EntoJoe]

Hi James,

At the very beginning of the run, there is only one ion being detected (m/z 79) and when it drops off there are no ions being detected in the baseline and very few ions (less than 5) being detected in any peaks.

I have run both an autotune and a tune evaluation, and it doesn't look great I don't think. I have uploaded the outputs from both of these operations to this dropbox folder, along with a couple of sample chromatograms (the torn and bruised folder is a sample that worked correctly about a month ago, while the BW03 sample is one I ran yesterday): https://www.dropbox.com/sh/lh2icb1bhg4i ... WRi6a?dl=0.

I have checked and there are no timed events set to run during my sample run from what I could tell, however I am not a GC-MS expert.

@cleh: I'm afraid I do not have access to the GC-MS until Monday morning now as the lab is locked on the weekends, but I shall post a spectrum as soon as possible.

Thank you again for your help,

Joe

[James_Ball]

Looks like you have very high air and water, especially water. There should be almost no masses present below m/z 50 in the scan when tuning.

Try tightening the nut where the column enters the mass spec about 1/8 of a turn and see if there is any change. Also check to make sure your Helium tank is not down to the last 100psi, because when it gets that low you are sucking up any impurities that may have settled to the bottom of the tank, including moisture. After that, if it does not improve then you will have to begin to look at any gas purifier filters that are in the gas line and check all connections for leaks. Also have to check the septum and column fittings for leaks too.

The only interference I can come up with at m/z 79 would be bromine, hopefully someone has not been using that on the instrument.

[Ian_R]

The only interference I can come up with at m/z 79 would be bromine, hopefully someone has not been using that on the instrument.

Br+ will give 2 peaks of equal intensity at 79 and 81. Br2+ will give peaks at 158/160/162 in a 1:2:1 ratio.

Not sure what 79 might be.

[James_Ball]

I was looking again at the autotune report and noticed you have a column 1 flow at 0.31ml/min and a column 2 flow at 1.45ml/min. Are you running dual columns combined just before entering the mass spec? (I have one set up that way and it is very tricky to get them right).

If so, check for leaks at the union to make sure that is not where the air and water are entering the system. You will need an electronic gas leak detector, don't use soapy water as that will be drawn into the instrument. If you must use liquid immerse the connection in methanol and look for bubbles, that will be easier to bake out of the detector than water.

[EntoJoe]

I was looking again at the autotune report and noticed you have a column 1 flow at 0.31ml/min and a column 2 flow at 1.45ml/min. Are you running dual columns combined just before entering the mass spec? (I have one set up that way and it is very tricky to get them right).

If so, check for leaks at the union to make sure that is not where the air and water are entering the system. You will need an electronic gas leak detector, don't use soapy water as that will be drawn into the instrument. If you must use liquid immerse the connection in methanol and look for bubbles, that will be easier to bake out of the detector than water.

Hi James,

Sorry for the delayed response, but I have been rather unwell and wasn't being intentially rude - I really appreciate the help the users here have provided thus far.

Just to update you as to my situation regarding the baseline issues. During my absence from work the labs had their GC-FIDs serviced, and asked the engineer to have a quick look at the GC-MS whilst he was on site. He essentially seems to have confirmed that there was too much water in the system so he changed the external filter (which was saturated) and cleaned the ion source. This has reduced the water content to around 10%, however there seems to be a high proportion of nitrogen in the autotune results now (35%) so I'm still having issues.

As for the columns, I am running two columns (both HP-5MS) as the front inlet is solely used for liquid injections/SPME while the back inlet is permanently connected to a Markes UNITY 2 thermodesorption unit. I shall have a look at the connection tomorrow with the leak detector.

Again, thanks for your help.

Joe

[James_Ball]

If you have the two columns installed into the mass spec with a Y connector, make sure the transfer line going into the mass spec is a 0.53 ID column and not a 0.1 ID. I tried the smaller transfer line thinking it would essentially serve as a no-vent type fitting to allow changing columns without venting but I discovered problems with getting proper flow through the columns. It essentially made the pressure/flow profile of my 0.25ID columns respond like they were 0.18ID columns.

You need the full vacuum at the end of the columns for proper flow to be delivered by the instrument as calculated from the head pressure. Unless you can somehow measure the pressure in the union, you can't get the instrument to calculate the proper head pressure to deliver the desired column flow. Once I went to the larger diameter transfer line, my analyte retention times were dead on what they were supposed to be versus reference chromatograms and dead times measured by injection nitrogen were perfect with what was calculated. Also when I was using the smaller diameter transfer line, I had an undulating baseline during the temperature ramp, which I think was from the two columns having flow pulses as the combined flow was too much to be pushed through the transfer line. It gave symptoms that looked like very high column bleed, but once I got the flows right it went away.