Tailing - chlorinated aromatics with phenol and aldehyde

[khirth]

Using new Varian 4000 GC-MS with CTC Combi-pal autosampler. Trap components are Silchrom-coated as is the 3800 GC injector. (Also tailed with old Thermo ion trap without Silchrom.) Liner is Siltek-coated single gooseneck with carbofrit. Have tried 3 different manufacturer's 5ms columns and all tailed. Now trying a VF-200ms column, but still see tailing. Grob test mix looks as follows: 2,3-butanediol has ~30% response of pentane, 2-ethylhexanoic acid also has decreased response with some tailing, octyl- & decylamines tail terribly but are visible, nicotine and dicycloamine show a bit of tailing, but notably nonanal is barely detected; other grob components have good response and shape and separation. My samples are in acetone and have been cleaned up by HPLC fraction collection and solid phase concentration. Analytes are aromatic ring with chlorine, phenol, aldehyde and methoxy groups. Doing quantitation of trace amounts. Using splitless injection with pressure pulse on 280C injector, fast injection, and oven temp starting at 40 because of low boiling acetone. Having a dickens of a time with tailing on these analytes. Have never deactivated a syringe - don't even know if it is a silly thought. If this is a reasonable thing, how to do or suggestions to purchase? Any other thoughts? Any and all suggestions to improve peak shapes of these analytes would be greatly appreciated.

[Peter Apps]

By far the most likely culprit is the carbofrit. If you are doing splitless injections you do not need flash vaporization so it is better to use an empty inlet liner - I would recommend a double gooseneck.

Deactivating the syringe will not make any difference.

Be sure that your carrier gas if free of water and oxygen - you can make a column active in one temperature cycle if not.

Regards Peter

[khirth]

Thanks so much for reply! Forgot to mention in original post that the carbofrit is a new attempt. i.e. tailed before when using single gooseneck with Siltek glass wool and also with empty deactivated glass double gooseneck. Carier gas is ultra high purity helium with Supelco high capacity (heated) gas purifier (cartridge changed fairly recently) + Varian's gas clean filter. (I'd be incredibly surprised if the culprit is the carrier gas.) To date I've not tried a Siltek double gooseneck - only deactivated glass - so I'll try that next. I'll also immediately remove the carbofrit since not necessary with splitless. Thanks! Anymore thoughts?

[JI2002]

A few things come to mind: First, try a different solvent like methylene chloride and see if tailing improves. Second, double-check to make sure splitless hold time is not too long, depending injection volume and liner dimention, it should be less than 1 min. Sometimes it can be as little as 0.1-0.2 min. Third, check split vent flow after vent valve opens up and make sure you have enough vent flow to purge out anything in the liner quickly after injection.

[khirth]

Tried a few things:

Changed liner to deactivated glass double gooseneck (don't have siltek in that shape).

For kicks, clipped head of this new column while injector was cool to change liner.

Decreased splitless time from 0.5 min to 0.1 and 0.3 min. Observed almost nothing at shorter times, so increased back to 0.5 min. Maybe this is diagnostic???

Checked split vent flow. Measured ~165 ml/min during run. Split is 100. Pressure pulse is just slightly longer ( 0.55) than splitless time to help ensure good purge of injector.

Watched EFC display for column flow. Observed steady 1.0 ml/min with steadily increasing head pressure. Seems in order.

Made new standard with all analytes using MeCl2 instead of acetone. Still observed gross tailing. (Interestingly, the non-chlorinated internal standard has even poorer response in MeCl2.) (Aside, have no control over samples coming in acetone. Want to avoid MeCl2 because need to prove biological source of chlorine, as opposed to lab handling.)

Rechecked Grob TestMix to ensure column not inadvertently trashed. TestMix is still same as before, so do not expect bad column and have good ion trap response.

Carrier gas is ultra high purity with both the Supleco high capacity purifier and the Varian pruifier plumbed in.

Any other thoughts, please?

Also, the oven temp program starts increasing immediately. Should I have an initial hold that matches the splitless time?

[chromatographer1]

First always try to trouble shoot systematically.

You need to determine if the tailing is coming from the injection procedure, the column itself, or the detector.

Do you get tailing peaks when you do a split injection instead of splitless? If you do then no matter what you do with the splitless injection you will get tailing.

Proceed from there.

best wishes,

Rod

[JI2002]

It's a good idea to hold the starting temperature for a couple of minutes to ensure sufficient sample focusing at the head of the column. Also make sure all the heated zones are set properly. Don't know if it's possible for you to inject the analytes one at a time.

[khirth]

Checked if analytes tailed with split (25:1) also. Yes, they did.

Lowered injector temp from 280 to 220, but still tailed badly.

So... I vented system to clip the MS end of column as well. (Head was clipped end of last week.)

After pump down I will re-check with my standard containing all the analytes, as well as with full Grob Mix. I will incorporate advice to hold the initial temp a minute or so to allow focussing and post outcome.
Thanks for the feedback. It is appreciated.

[Peter Apps]

You have eliminated the obvious likely causes, so now we move to the obscure and unexpected !

You are doing a splitless injection with a low initial oven temperature so you might be getting some solvent recondensing on the column and causing peak distortions that look like adsorptive tailing. This could explain why some of the Grob components are tailed and some not, but cannot account for the problem still being present when you do a split injection.

Pressure pulses can send gasses in unexpected directions into undesirable places - like into gas lines and unswept areas. Try an injection with constant inlet pressure.

There may be some crud sitting in your inlet body below the liner. In principle this region is swept to waste by the split purge flow, but it does not always work out so simply - the Gerstel CIS inlets are very vulnerable to this and Agilent inlets accomodate it by changing the gold seal. With the Varian you will need to separate the bottom part of the inlet from the rest of the body and check for bits of ferrule or whatever.

The Varian 3800 has a commendably simple septum purge control, so simple that it is easy to neglect it. Check that you have about 4 ml/min coming out through it.

Try doing a hot needle injection by holding the syringe in the inlet for 2 s before and after you inject.

Good luck

Peter

[khirth]

Using brand new Varian 4000 ion trap with 3800 GC.
Analyzing for aromatic adehydes with chlorine and phenol functionality.
Have tried:

3 different manufacturer's 5% phenyl - all tail.
Purchased VF-200ms (marketed for electron rich halogenated aromatics and aldehydes) - no better.

Using ultra high purity helium with both Supelco high capacity air purifier and Varian's purifier.

Tried several inserts: deactivated glass double gooseneck; deactivated glass single gooseneck with wool; Siltek single gooseneck with and without carbofrit and with Siltek wool - all tail badly. Currently using empty Siltek single gooseneck.

Observe severe tailing in both splitless mode with pressure pulse and split mode constant pressure.

Checked all flows:
Measured split vent flow ~165 ml/min mid-run with 100:1 split.
EFC reads steady 1.0 ml/min with column head pressure rising steadily throughout run.
Measured septum purge is 3.0 ml/min (as set up at install).

Lowered injector temp from 280 to 220 - no improvement.

Have cooled everything and clipped both ends of column - no improvement.

Grob mix is still running same as when column (VF-400ms) was new: 4-Cl-phenol runs fine; nonanal very small and broad; amines, of course, tail badly as does 2-ethylhexanoic acid.

Have baked out trap with no improvement. Trap response is good on Grob test mix. Really don't think it's a trap issue.

Made up new standards in MeCl2. Same tailing as observed when made up in acetone. Don't think it's an acetone issue.

This morning, as result of more advice at chromforum.com (Thanks!!), tried hot needle injection (2 sec pre-injection delay). Saw absolutely nothing of the Standard mix! Then injected Grob mix, which ran fine.

Concurrently (unfortunately) daily check indicated high water so vented, replaced both ferrules after clipping both ends of column again.

Also cleaned injector (1177 constant temp) body as best I could. Swabbed inside with MeCl2/MeOH. Visually inspected using flashlight and saw no bits of septum or anything (but truly couldn't see much). Swabbed from bottom to try to rid any graphite residue, then pushed a clipped piece of column through. I really don't know what to do if 'bits' get into the bottom of this injector. Constructive thoughts/suggestions?

Another wild idea I had is to flood the injector with several back-to-back splitless injections of a non-target analyte of similar characteristics to "coat" the active surface of the injector, then trying my samples...but I don't want to exacerbate an already big headache. Thoughts?

Again, thanks for taking the time to offer some good suggestions. It is much appreciated.

[Peter Apps]

Many, many years ago the wise old men of gas chromatography said that you should never use acetone as a sample solvent on a column with a silicone stationary phase. Nobody seemed to know why, and it did not make much sense to me.

I wonder if that is not what has gone wrong here. Maybe the folk lore is right. You get the same results with methylene chloride but that might be because the acetone has already damaged all three columns.

Could you try a PEG phase column ?

Peter

[chromatographer1]

Personally I would like to see the sample and the columns used on another detector. While acetone may very well be a problem with the surface chemistry (especially trace amounts of impurities in the acetone) I would want the detector to be eliminated as a cause of the tailing.

I hope this helps

Rod

[JI2002]

Because you have both acids and bases in the mixture, there could be some interaction between them in the injector. I would suggest to inject the analytes individually to rule out the possibility.

[khirth]

Few updates:
Tried injecting stdards indiviually - still tail badly

Injected a non-chlorinated similar as a single componenet - tailed badly

Cannot use different detector as don't have one. (I don't think the trap is the problem though.)

Contacted Supelco tech service: was assured with emphasis that modern columns will not suffer from acetone as solvent, though that may have been the case long ago.

Need to mention that my solvent is the highest quality I know of: B&J GC2. So I don't know about impurities in the solvent, although I'll keep that in mind even though I'm stuck with acetone. If you have comments on other solvents I'd appreciate it.

Installed SupelcoWax10. I'm guessing the Peter Apps suggested this phase because it is so completely different than the others and should, therefore, be more rugged to acetone.

Have also prepared a series of non-chlorinated similars:
Vanillin (4-OH-3-OMe-benzaldehyde) = aromatic phenol + aldehyde
4-OH-3-OMe-benzyl alcohol = aromatic phenol + alcohol
4-OEt-3-OMe-benzaldehyde = aromatic aldehyde
4-OMe-benzyl alcohol = aromatic alcohol
3,5-di-tert-butyl-4-OH-benzyl alcohol = aromatic alcohol + restricted phenol
3',5'-dimethoxy-4-OH-acetophenone = aromatic phenol + ketone
3,5-dichlorophenol = chlorinated aromatic phenol (ortho as are my analytes. Grob 4-Cl-phenol is para.)
3,4-dimehtoxyphenyl acetone = ketone with aromatic substituent

Will run these on Wax10 this week-end and post outcome Monday.

P.S. Still using Siltek single gooseneck with no packing.
Supelco tech service is going to talk among themselves and then send me a different insert. I appreciate this too, as all the constructive suggestions from this site.

[khirth]

Without changing anything more in the injector and without changing any parameters in the trap, I installed a Supelco Wax10 column, I injected these:
Vanillin (4-OH-3-OMe-benzaldehyde) = aromatic phenol + aldehyde
4-OH-3-OMe-benzyl alcohol = aromatic phenol + alcohol
4-OEt-3-OMe-benzaldehyde = aromatic aldehyde
4-OMe-benzyl alcohol = aromatic alcohol
3,5-di-tert-butyl-4-OH-benzyl alcohol = aromatic alcohol + restricted phenol
3',5'-dimethoxy-4-OH-acetophenone = aromatic phenol + ketone
3,5-dichlorophenol = chlorinated aromatic phenol (ortho as are my analytes. Grob 4-Cl-phenol is para.)
3,4-dimehtoxyphenyl acetone = ketone with aromatic substituent

Observed great peak shapes on all (except saw nothing with 4-OH-3-OMe-benzyl alcohol, but will troubleshoot that later).

Then set up 10 point standards bracketing about dozen samples and got very respectable results. Acetone rinses were injected between each and every sample and standard to prove no carry-over. The peak shapes remained good throughtout and the slope increased (not decreased!) slightly with the end bracket.

Peter Apps suggested go to WAX column. Thanks!


Tonight I am running a very, very long sequence with acetone as solvent. I will report back
as to how the WAX column holds up.

QUESTION: Does anyone out there know WHY acetone has this effect on the 5ms, and 200ms columns?

ACTION: I'm adding some of these chlorinated analytes to my Grob mix.

QUESTION: Are there other solvents that experience says should try to steer clear of with bonded phases? (Is that the correct term: bonded phases?)

QUESTION: Anyone have a good, concise explanation as to the intrinsic differences of the bonding of the other phases and WAX? Or why the WAX is not appearing to be subject to the acetone deterioration?