Many regular ghost peaks with m/z 40 in first few minutes

[mphillips]

Greetings,

I am experiencing a series of ghost peaks early in my GCMS runs where many very sharp peaks with m/z40 appear in the first few minutes of each analysis. These peaks are not related to the sample and occur in air blanks as well. m/z 40 could be argon but I don't know where this would come from or why it would show up as a series of ultra sharp peaks early in each run. Any suggestions?

Thanks,
Mike

[Don_Hilton]

A bit mroe information may help get an answer:

How large a the peaks - height and width. And since scale varies from instrument to instrument - what kind of instrument are you running (make, model and software version)? And some problems tend to be more common in particular instruments.

For multiple sharp peaks, a picture might help. There are instructions on the site for posting pictures.

What is the dead time for the column as you are running it? And, what time do the peaks start? What are your injections conditons - in particular, injection mode, column head pressure and valve timing? Are you using gas saver and what condition is the inlet before befor the GC is to go to ready for an injection. Does the GC run constant flow or constant pressure? Knowing the GC oven profile would help to match up temperature changes with what is going on in the column.

It coudl be as simple as a ferrule that reseats at a particular oven temperature - so, you have checked to be sure that all column and inlet connections are secure? Septum has been changed reciently?

Mass 40... What mass range are you looking at? Does m/z 44 show similar spikes, perhaps a bit rounded and delayed from those for m/z 40? (CO2 can be retained a bit, even on something like a DB-5.) I assume this is EI? How does your air/water leak check look? If the instrument is tight enought, picking up data for m/z 32 or eve 28 on an injection might be helpful.

[AICMM]

You are scanning from what to what? Probably 35 and up? Try looking at the tune for air leak. Argon is 40 and 1% in the atmosphere. More info better answers.

Best regards,

aICMM

[mphillips]

Hello again,

First of all, thanks for the speedy reply. I think I have found the source of the problem, but in case it comes up again, here are (most of) the answers to your questions:

1. The peaks are very sharp, only a single scan wide it seems, and regularly spaced, although right after the solvent delay it seems like a high background, breaking down into the forest of sharp peaks with a width of about 0.01 min.
2. I am running a Agilent 7890A and using ChemStation version E02.02.1431.
3. The dead time of the column is about 2 min.
4. The peaks start right after the solvent delay, which is 6.4 min in this method--long but necessary to elute unreacted derivitizing reagent. The ghost peaks are present from this moment until about 9 min.
5. Injection conditions are as follows: split 1:10, 9 psi column head pressure, constant flow 1 mL/min, injector port at 250 C, and the method uses gas saver when not running.
6. You asked about the condition of the inlet before the injection but I am not sure what you mean.
7. The oven profile is 70 C x 1 min, then 1 C/min to 76, 5 C/min to 220, 12 C/min to 320, and hold for 8 min.
8. The ionization is electron impact
9. The air&water check is at an acceptable level (<5% of the PFTBA calibratin standard m/z 69 peak), but only after changing ferrules.
10. Scanning from 40 up.

I think the first reply was right about a loose ferrule. I changed the inlet liner, septum, and ferrule on the injection side. I also located a leaky Ar tank that was bleeding Ar into the lab air space. The Ar was probably showing up in every analysis because the GC was not completely sealed. I am not sure why it showed up as a forest of peaks at the beginning and not continuously. Perhaps someone has an explanation for that or a better explanation for why these peaks appear. I have now obtained a clean blank run after changing ferrules, septa, etc., but it could always crop up again if I have not diagnosed the problem correctly, so still looking forward to your comments if you have any.

Thanks in advance,
mike

[Don_Hilton]

My suspicion is that those spikes were the smallest unit in height you can get from the detector? If so, what you are seeing is either from the tail of the solvent peak or an air leak being sealed as the oven heats. The signal is just at the point that it is detected with the lowest count rate that can come from the detector or not detected - or ocassionaly twice the lowest rate, depending on electronic noise in each scan. As the leak or tail finally goes away, the count becomes consistantly zero. Depending on solvent you could have m/z 40 present, but more commonly m/z 40 is from air in the room. There is sufficient argon in the air to be noticible with even a small leak. The reason I asked about m/z 44 was to see if you saw this signal as well - CO2 is generally a stronger signal in air.

[Darber99]

m/z 40 is usually a sign of a thermal leak. My guess is you have m/z 28 and 18 (Nitrogen and water) above 10% of m/z 69 as well. Try giving your source nut a 1/4 turn (but be careful not to break your column!) and see if those ions disappear. Good luck!