Chromatography Forum

Dear people, I have this problem with my GC/MS (3900/2100T) system

I have created small webpage with screenshots which should explain everything:

http://www.ktf-split.hr/~franko/eug/index.html

Why are these peaks here?

Diethylphthalate could be from the solvent or due to the sample preparation?

Have you tried another batch of pentane?

To sharpen the solvent peak start from 50 C and you should already see a difference.

I'm not familiar with Varian instruments but 2 to 4 ul might be too much for your liner (you have to make sure that you don't exceed its volume with the ul of vaporized solvent injected).

Diethyl phthalate peak was present in three different batches of solvent i think two at least were from different company (pentane solvent)...?

And these 2 and 4 uL injections were just a test for split modes, peak appeared the same way when we injected 1uL and 1.5uL, so it might be ruled out, and even if it is so how can it again exit after 60min (out which last 20min is at 220°C) when column oven is cooled down... ?

We use to have old hp5980 gc and 5971A MS i think and when we would put 3min solvent delay there was no solvent peak...

I'll try with this 50°C temp, though, maybe that helps, although I would be most happy if it wouldn't be there at all along with phthalate peak.

Is there a chance that you extract this phthalate during the sample preparation? Pipette, vial or vial's septa?

If you concentrate 20 ml of pentane to 1 ml or less, do you observe a higher signal for this phthalate?

What kind of liner are you using? Glass wool, baffled?

Do you see an abnormal quantity of solvent as background ions in the baseline (when extracting masses 57,71)?

If you inject another solvent do you still see the phthalate?

Last question of the day...

As I wrote I'm not familiar with Varians but is it normal that you have a split 1/50 and that at 0.01 min it goes off (see the 2nd printscreen of your website)?

On your splitless pentane injection, how long is your split closed before venting open to rid the injector of excess solvent? I would suggest a splitless time of 0.75 minutes, then open the split at 50-100:1 at 0.75 min to vent.

As for the diethyl phthalate, since it is so well resolved from your eugenol & a-pinene, does it really matter? It's a pretty common contaminant.

First step is to increase the splitless time - that 0.01 min will cause all kinds of gas flow peturbations. In fact, with a 0.01 min splitless time you are effectively working split, so unless you need detection at very low levels a simple split injection, say 10 or 20:1 might be the best option. If that does not help then remove the inlet liner and look down into the inlet body - you might find some septum fragments or other dirt lying at the bottom which do not get removed when liners are changed. The easiest way to remove them is to put a glass pasteur pipette on the end of a plastic pipe and suck them up with a vacuum.

The phthallate could be coming from anywhere. Unless it is interfering it is good advice not to worry about it.

Peter

leoglag,

I may be reading this wrong, but I don't think you ever go back to split. At 0.01 you go to splitless and then you never switch the valve back to split. This would explain the really broad pentane peak out at 6-10 minutes since you never switch back to split and clean out the injection port. I would add an event at 0.5 min split state on at 50:1.

Best regards.

Thanks for all of your suggestions.

This what I did:

I've put 3 events:

1. Initial Split is Off
2. 0.7 min Split is On 1:50
3. 3 min Split is On 1:20

At these settings, solvent peak (1. peak, big one, ) disappears to the baseline and it is almost always up to 50 kCounts high totally invisible!

And Diethyl phthalate peak is also relatively invisible - and I decided to ignore it; like You said - it is a common contaminant...

So that's it. I guess it will be good now.

But there is another problem now concerning report generation - I'll make another topic for it.

Anyway thank you all.

Thanks for the feedback - glad it worked.

Peter