GC carry over or anything else?

Chromatography Forum: GC Archives: GC carry over or anything else?
Top of pagePrevious messageNext messageBottom of pageLink to this message  By Anonymous on Tuesday, December 4, 2001 - 12:09 pm:

Hi, I am wondering if anybody could help me out of the following problem.

Currently I am ruuning 60 plasma samples with GC/MSD. I set up my sequence in such manner that every working standard is followed by a blank (in case of carry over)and then 6 samples. So I have 10 subsequences totally, and there is a ws and blank vials for each subsequence. All the ws, blanks and samples are injected in duplicate.

The first 4 blanks gave minimum peaks for the target compound and the internal standard. Then suddenly from the 5th blank to the 10 th blanks, the 1st injection of the blank generated peaks for the target compound and the internal standard, the peak area for both compounds are about 3% of the ws. The peaks in the second injection lowered to about 1/8 of the 1 st injection.

My GC/MS condition is:


Oven Initial temp.: 150 C Initial time: 4 min Ramp: 20 C/min Final temp.: 300 C Final time: 5 min Total run time: 16.50 min
Inlet Mode: Split Split ratio: 20:1 Initial temperature 250 C Gas saver: on saver flow: 20 mL/min Saver time: 2 min Carrier gas: Helium Column: Constant flow (0.5 ml/min)

MSD Transfer line Temp.: 280 C MS Quad. Temp.: 150 C Ion source: Electron ionization SIM

So my question is what factors could possibly cause the followed problem, and how to correct it.

I would appreicate any body's input and thanks in advance.


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Jason Ellis on Friday, December 7, 2001 - 11:59 am:

If you are getting peaks then we know it has to be coming from somewhere in front of the analytical column (i.e. in your inlet, syringe or vials). You didn't say how much volume you are injecting -- what is that number? Also, what is your solvent and what are your compounds of interest? You may be backflashing in the inlet and condensing analyte on cold spots. This condensed analyte can later be "rinsed" back down with subsequent injections to give peaks in blank runs.

Do you get any peaks if you let the instrument sit idle (oven at cool temperature) for a while (i.e. overnight)? Try letting the instrument sit like this then do an instrument blank -- just hit start on the GC and acquire the data, don't inject anything though. If peaks come off during this run then we can be pretty sure you have a contaminated inlet, probably due to backflash. If this is the case, cool the inlet down, uninstall the column and liner and physically rinse the inlet with solvent to remove the contaminant material. Also a good idea to change septum, liner and gold inlet seal (if it's an Agilent/HP instrument).

Hope this helps. Let me know if you have any questions.

Regards,
Jason Ellis
Technical Support Engineer
Agilent Technologies


Top of pagePrevious messageNext messageBottom of pageLink to this message  By bill on Monday, December 10, 2001 - 04:55 am:

Two suggestions:
1- It sounds like the wash vials have gone empty.

2- I would recommend instead of running empty vials to check for carryover. Run Solvent with internal standard. That way you can directly compare and quantitate the carryover.


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Anonymous on Monday, December 10, 2001 - 11:23 am:

Hi, Jason and Bill,

Thanks a lot for your input!

According to Jason's suggestion, I let the instrument sit idle over the weekend, and started it three times without any injection this morning, it turned out that no peak showed up. Does it indicate the inlet port is safe?

Sorry, I missed the injection volume in the original message. The injection vol. is 1 ul, and the silvent is ethanol. Both internal and target compounds are basic drugs.

Bill, I was suspecting wash solvent(ethanol)too. Actually, I did a couple of tests on the wash solvent last week (4 days after the sequence was doine)in the following manner

sequence: 3 solvent vials followed by one working standard and two solvent vials. singlet injection

test 1: use the same wash solvents left over from the sample assy

test 2: new solvent

test 3: old solvent from the sample assay

test 4: the solvent used in test 2, also made a injection of the old solvent from the sample assay

test 5: switch to MeCl2 as wash solvent, also made an injection of the old solvent

Results (peak height)

Test 1: 1300 (vial1); 100(2); 30(3); 160000(4,ws); 2600 (5); 300 (6)
Test 2: 20 (1); 20 (2); 10 (3); 170000 (4,ws);
70 (5); 30 (6)
Test 3: 50 (1); 50 (2); 40 (3); 160000 (4, ws) 2200 (5); 220 (6)
Test 4: 340 (1); 100 (2); 30 (3); 160000 (4,ws); 2200 (5); 400 (6); 80 (7, old solvent)
Test 5: 150 (1); 40 (2); 10 (3); 160000 (4,ws); 300 (5); 240 (6); 90 (7, old solvent)

Based on the blank injections suggested by Jason and these test results, it seems the problems lies in the wash solvent or injector (if I am right), but I still could not locate the casue of the problem.

I would appreciate any one's generous help!


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Bruce Freeman on Monday, December 10, 2001 - 12:19 pm:

First of all, Jason is right on the money with his suggestions, and sounds to be slightly more current than I am (which is not surprizing since he works for Agilent).

To recapitulate, the problem is almost certainly ahead of the column. (I'm sufficiently unsure about the GC-MSD interface that I don't want to proclaim that a certainty.)

Since you're only injecting 1 mcL of ethanol, and you're running in split mode (high flow through inlet) you're not likely to be seeing the results of flash-back. That happens more in splitless injection, especially where there is too large an injection of a light solvent (e.g., water) and/or with an inlet liner that has an especially low volume or a restriction of some sort. I'd like to know what your inlet liner is, but I discount this mechanism as a likely problem.

Backing up a step, I recommend to anyone using an Agilent GC to rebuild the inlet whenever you start up work on the GC. It's easy, and not too expensive. Consider septum, inlet liner, o-rings, syringe and (to a lesser degree) gold seal (at bottom of inlet) to be disposables. Keep them in stock and change them "frequently."

People forget that a syringe wears out fairly quickly. Examine your syringe, up near the top of the barrel (i.e., away from the needle). If there's a ring of black about 5 mm long about 5-10 mm below the top, throw away the syringe. That's metal worn off the plunger. You can't save the syringe. It leaks. It's trash.

Always use solvent wash and remember that the needle does not go all the way down to the bottom ofthe vial, so the bottom few mL are "wasted". HP used to include a vial with a mark to indicate the bottom of the useful "fill;" I don't know whether Agilent still does.

What I'm driving at is that it's a great deal easier to prevent problems on an Agilent GC than to diagnose and cure them after they develop.

One further thing that caught my eye: What exactly is the form of your sample? You say it's "plasma." I take it you mean blood plasma, presumably diluted? Since blood plasma contains considerable solid material, that material has to go somewhere. Presumably you're catching it on glass wool or packing in your inlet liner. What does that liner look like after a run? Crud in a liner can cause a lot of problems. It can change the evaporation characteristics of the sample. It can lead to carry-over by "encapsulating" analyte from one injection and releasing it when the next solvent injection hits it. Always be suspicious of an assay in which the inlet liner is used to filter crud from the sample. (None other than Walter Jennings taught me that one, more years ago than I care to admit!)


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Jason Ellis on Monday, December 10, 2001 - 03:48 pm:

I'm with Bruce on his suggestions. It's probably not backflash as I first suggested -- it would be difficult to get 1 uL of EtOH in split mode to backflash. I'm not sure if I completely understand the results you posted, but it does appear that you're getting some carryover from your solvent rinse vials -- especially for peak #5. What are your autosampler parameters (i.e. how many solvent A/B rinses do you do before and after, and how many sample rinses do you perform prior to injection)?

Agilent still sells the marked rinse vials for the 7673 autosampler. They have a "minimum solvent" mark printed on the vial. E-mail me privately for the part number if you'd like it. I don't want to be giving out P/N's in this discussion group -- too "commercial"! :)


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Anonymous on Wednesday, December 12, 2001 - 03:07 pm:

Hi, Bruce and Jason,

Thanks a lot for your great help!

Following please see my autosampler parameters:

Preinjection wash

Solvent A (ethanol) 3 times
Solvent B (ethanol) 3 times
sample 2 pump strokes

There is no post-injection washes.

For Jason, what do you exactly mean "carry over from solvent rinse vials"? you mean the rinse solvent got contaminated or anything else?

If the rinse solvent got contaminated, then the injection of this solvent should generate a peak, but it did not (please see the results in test 4 and 5, peak 7). am I right?

As for syringe, it did get wear off, but it does not like the main factor becasue after I installed a brand new syringe, the carry over was still there.

It seems I will continue to bug you guys for some time.

Thanks a lot!


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Jason Ellis on Thursday, December 13, 2001 - 08:34 am:

I meant that you might be getting carryover of your target analytes from the solvent rinse vials. It appears to me that you get an increase in all peaks (especially #5) in the "used" solvent vials when compared against run #2 (fresh solvent).

You might try adding a similar post-injection rinse scheme (i.e. 3 rinses in each vial). You might also try adding a few "sample rinses" to your method -- where the autosampler pulls back a syringe load of your sample and deposits it to waste. This helps to rinse your syringe with the actual sample you're going to inject. I'd try adding 2 of these to the method.


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Anonymous on Wednesday, January 2, 2002 - 03:20 pm:

It turned out that the trouble-maker was the wash vial. The wash solvent gradually went below the "minimum" mark during the sequence ruuning. I was not using the Agilent vials at that time and was not aware of the mark.

Sorry for the late update.

Thanks a lot for all your help and happy new year!


Posting is currently disabled in this topic. Contact your discussion moderator for more information.