more on tailing, acetone, non-Wax phases --> my analytes

[khirth]

First, I apologize if this is getting to be redundant for some, but this has been a real challenge for me. And, I still don't know what's going on...

To review:
My analytes were phenolic aldehydes which have been chlorinated by a fungus. i.e. aromatic ring, OH, CHO, OMe, Cl and monomeric.

My samples have been cleaned up by SPE and eluted with acetone, which is a very good solvent for lignin and the like. I have no choice over the solvent.

I started the GC using a ZB-5ms column. First few runs looked OK, but then tailing developed and worsened horribly. Tried 2 other manufacturers columns (Supelco and Varian Factor Four) which also exhibited severe tailing. Got so bad there was almost no discernable peak from the standards.

Switched to VF-200ms column marketed for use with electron-rich compounds like halogenated aldehydes. Still very bad tailing.

Started looking very closely for instrumental sources:
Verified split flow and septum purge, clean inlet, ultra-high purity helium with 2 different in-line high-capacity gas purfiers with color indicators.

Removed all packing from liners. Tried both deactivated glass and Siltek-coated single gooseneck liner. Same result: bad tailing.

Ion trap is clean as a whistle.

Removed pressure pulse from splitless injection.

Did split injection and still observed bad tailing.

Lowered injector temp from 280 to 220. No improvement.

Raised initial column temp from 40 to 140. No improvement.

Clipped head of column, then clipped head & tail.

Using highest quality acetone I know (B&J GC2).

Made up stds in MeCl2 (B&J GC2) and still saw bad tailing.

Made up indiviudal stds in MeCl2 to ensure no interaction between the target compounds in the inlet. Still tailed badly.

Speculation was that the acetone had ruined the column so making the stds up in MeCl2 had no effect because of the previous acetone.

NOTE: Grob Mix still looked same as new: tailing diol & amines; 4-Cl-phenol was OK, nonanal was not OK and showed presence of nonanoic acid. But since this had not changed since the column was new, I assumed the column was still OK.

Injected a series of similar compounds which did not contain all of the functionality of my analytes. Other than the 3,5-t-butyl-4-OH-benzyl alcohol, all showed very nice peak shape and they were in acetone.

Finally, followed suggestion by Peter Apps to try a Wax column. That did the trick!

Ran very long sequence with multiple samples, bracketed by 10-point standard curve, including 5 QCs within the sample sequence and each and every sample/std/QC was followed by at least 1 acetone rinse to ensure no carry-over (to prove source of chlorination). (Chlorinated aldehydes give very pretty pattern of M+, M+2, M+-1, M+-1+2.) Beautiful results, very respectable linearity and sustained response.


Then it occurred to me that while the Wax column worked just fine with the acetone solvent, I had not truly proved that acetone is incompatible with the bonded phase column. i.e. was it the acetone or was my analytes?


The only brand-spanking new column I have is a ZB-1ms.
With utmost care (i.e. cool injector, oven and transfer line; continued flow) I installed the brand new ZB-1ms not connecting it to the MS. Swabbed the injector and installed new siltek single gooseneck liner. After flowing 15 min with high split to purge injector, I conditioned the column 40C for 15 min, to 300C at 5C/min, hold 30 min. I cooled everything again, connected to MS, and conditioned again. I did everything I could to ensure no oxygen saw this column while it was hot.

Then I injected a standard mixture in MeCl2.
On this column that has never seen acetone, my analytes tail and have significantly decreased response. The non-chlorinated internal standard (4-OMe-cinnamaldehyde) does not tail. Whereas, on the Wax column this standard gave approximately same response for all chlorinated analytes and the internal std with all having good peak shape.

So, yesterday I made up another batch of the similars only I used MeCl2. the similars are: 3-OMe-4-OH-benzaldehyde (vanillin), 3,4-OMe-phenyl acetone, 3,5-chlorophenol, 4-OH-3-OMe-benzyl alcohol, 4-OEt-3-OMe-benzaldehyde, 4-OMe-benzyl alcohol, 3,5-OMe-4-OH-acetophenone, and 3,5-t-butyl-4-OH-benzyl alcohol. All, except the last which gave low response, gave excellent peak shape.

---> The similars which have some, but not all, of the analyte functionality do NOT tail. Whereas the analytes DO tail - even in MeCl2 - and using a brand new column.


---> Now I'm not so sure the culprit is the acetone.
The culprit may be the analytes themselves.


QUESTIONS: Have you ever seen something similar? Any other thoughts?

[chromatographer1]

I suspect because you have found that the analytes in question tail at the concentrations at which you are injecting them that the 'problem' lies in the surface chemistry of the columns that have a non-PEG phase coating.

There may be very little that you can do. Because of the emphasis on low bleed performance on silicone based phases in recent years it very well may be that the deactivation and phase chemistry to achieve these very low bleeds like the Factor Four of Varian and the ZB-5MS product lines are giving you the tailing that you are seeing in your analysis.

The different chemistry of the PEG type phases is giving you better peak symmetry performance. Why not just go ahead and use the PEG phases?

To improve (reduce) the tailing can you go to a thicker film silicone column which may give you the selectivity you prefer? Is using the PEG column a problem?

I have seen the effects you describe on our competitor's columns and I believe it is due to surface chemistry issues. I have not tested our new low bleed columns with these analytes (I work for Supelco), but that is not within the work I normally perform here so I cannot offer any proof statements.

Hope this is helpful.

Rod

[khirth]

Dear Chromatographer 1:

Indeed I do plan to continue using the Wax columns because they work - which, in the end, is what counts. It was just so bizarre I thought it was worth noting - especially given the fact that I had conflicting advice about acetone being the cause.

Also, as an fyi, the Supelco column also exhibited the same gross tailing as the Varian Factor Four and the Zebron. So, as it turns out, this phenomenon was independet of column manufacturer, which is one of the first variables I tried.

Also, Jerry, at Supelco's tech service, was very helpful and interested. Jerry was pretty adamant that acetone can be used on modern bonded columns with no ill efffects, although he said it was apparently a problem in years past.

Thanks for your help. P.S. I like your tech service.

[Peter Apps]

I have done chlorophenols on both silicone and ZBwax, and both gave sharp peaks, but with the ZBwax marginally better. The samples were injected in dichloromethane.

Adding chlorine to phenols makes them markedly more acidic, so the more chlorine they have the more they tail.

What is particularly puzzling is that aldehydes tail - even the nonanal in the Grob mix. I would expect carbonyls to be well behaved on pretty much any column.

What quantity of analyte are you putting onto the column ?

Just to be absolutely sure - when you say tailing you mean that the peak goes slowly back to baseline AFTER the peak maximum - not that it goes slowly UP to the maximum ?.

From what Chromatographer1 says about the new generation low bleed columns, this might be something that the column manufacturers should be taking a careful look at.

Best wishes Peter

[khirth]

Thanks for your attention and help Peter. It is very much appreciated.

First, by tailing I do mean the return to baseline after max hgt. Severely non-gaussian shape. To the point of making integration extremely iffy. Eventually after several injections leading to almost no discernable response.

Also, the chlorine does make the phenol more acidic.
But.... I don't believe it changes the pKa by more than 1 or 1.5.
pKa of phenol ~9, so it's just not that acidic.
(See http://www.chemicalforums.com/index.php ... eadid=1654)
If pKa is not the proper way to think of this, then please sing out because I am just learning here. (Anointment by fire, I think.)

The nonanal has significantly less response on all columns - new and used - of all phases. All show a large nonanoic acid peak also.
I will do an NMR to check my nonanal.

I read some reference article that describes the purpose of having nonanal in the mix is to test non-hydrogen bonding adsorption.
(www.quadrexcorp.com/new/grobtest.htm)
It specifies "saturated aldehyde," which is not what I have. Maybe significant.(?)


My Grob stocks are in hexane at:
Nonanal: 127 mg/10 ml
2,3-butanediol: 199 mg/10 ml
2,6-dimethylaniline: 106/10 ml
2-ethyl hexanoic acid: 120 mg/10 ml
methyl undecanoate: 121 mg/10 ml
etc.... (ranging from 78 mg acenaphthylene to 199 mg diol)
Grob mix is 1 mL each diluted to 200 ml with hexane.

As far as amount on column: I am using a splitless injection of 1 uL and seeing a range of ~0.3 to 13 ng depending on incubation time, medium and wood species. All the samples are bracketed within my 10-point standard curve, which has excellent linearity irrespective of analyte. (3 target analytes differ by position of chlorination and source lignin degradate. All have the aromaticity, aldehyde, phenol, methoxyl(s) and chlorine and are monomeric.) There is no indication of fronting, so I am assuming I am not overloading the column.

Have a good one. And, again, thank you very much!

[Peter Apps]

Hi Khirth

pKa is a handy proxy measure for how badly a peak is likely to tail because the lower the pKa the more polar is the phenol -OH group and the more tightly it sticks to -OH groups in the inlet and column. With a wax column this gives you retention, on a silicone column it gives you tailing.

pH is logarithmic, so a change of one unit in pKa implies peak shapes that are ten times worse.

Your details on the Grob mix help to clear up one mystery - why the Grob test chlorophenol was sharp, but the sample analytes were tailed. You are putting about 60 ng of each Grob component on the column, which is one or two orders of magnitude more than the analytes. With adsorptive tailing the bigger the peak the sharper it looks. If you dilute the Grob mix by 20 times or so you will probably see peak shapes similar to those form your analytes.

Your more dilute samples give sub-nanogram amounts of analyte on the column, doing chlorophenols at these levels is very tough, and you will probably always see some tailing. You could try increasing the injection volume, but that will open another can of worms.

If the wax column does the trick, go with that.

Peter