Reducing Split Flow vs. Splitless Injection

Chromatography Forum: GC Archives: Reducing Split Flow vs. Splitless Injection
Top of pagePrevious messageNext messageBottom of pageLink to this message  By Brett Libby on Wednesday, October 2, 2002 - 02:19 pm:

I’m currently using a Phenomenex ZB-Wax Capillary Column (30M x 0.32mm x 0.5micron) to separate ethylene oxide and ethylene chlorohydrin (in a water/methanol mix). My problem is I can’t get large enough peaks for my 1ppm samples with a 1 microliter injection.

The instrument Parameters are:
Helium carrier flow =2mL/min
Split Flow = 25mL/min
Injector and Detector = 250C
Oven = 60C
Temp Profile = 60C for 1.6min and then 12C/min to 200C

The peaks of importance are:
“Injection peak” = 1.47min (always the first peak)
Ethylene Oxide = 1.70min
Methanol (in solvent) = 2.43min
Acetonitrile (Internal Standard)= 3.42min
Ethylene Chlorohydrin = 8.23min

I’m attempting to increase the sensitivity of my method by decreasing the split flow or possibly moving to a splitless injection. I have a wide bore liner in my PE Autosystem GC instrument.

1. Decreasing Split Flow

As I decrease the split flow down to 5mL/min, the peaks early in the run are broadening and tailing such that the “injection peak” is running into the ethylene oxide peak and the methanol and internal standard(acetonitrile)peaks are tailing. How can this be prevented or minimized?

2. Moving to Splitless Injection

If I move to a splitless injection, what do I need to change or be aware of?

Do you have any other suggestions for improving the sensitivity (larger column ID? Longer column? Different kind of column?)

I’ve only been doing GC work for a couple months so I’m still learning a lot as I go. Any help would be appreciated. Also, if you need more info please let me know.

Thanks


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Anonymous on Thursday, October 3, 2002 - 06:03 am:

Do not try splitless unless you want to see an incredibly ugly chromatogram. You have a component eluting before one of your solvent components, and going to splitless will spread this peak dramatically. Try calling one of the column companies to see if they have information on this analysis.

Things I would try if I were doing this analysis would be narrower bore column to improve resolution and separation, lower starting temperature to try to improve separation, different temperature ramps and holds to try to impruve the separation.

As far as peak size is concerned, if the peaks are large enough to integrate reproducibly and give the required precision for the method they are good enough. You need to worry about precision and linearity, not just peak size.

Good Luck.


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Rodney George on Thursday, October 3, 2002 - 07:20 am:

I assume you are using FID. It is difficult to detect your analytes with a FID at 1ppm.

You may wish to try another detector. I have found a non-FID solution to your problem but it is proprietary information and I cannot tell you the details.

Good luck

Rodney George
Supelco Gas Separations Research


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Spencer on Thursday, October 3, 2002 - 08:34 pm:

You might try a megabore (0.53 mm) column with carrier velocities in the 30-40 cm/sec range - I can't remember the volumetric flow rate for this velocity with a megabore; either test with methane or butane (use a lighter and 100 uL syringe) or calculate it with FlowCalc or something. This would allow you to run at a 10-15 mL/min split flow and get decent peak shape (I wouldn't go below 10 mL/min on split applications, the instrument vendor may say 20). I disagree with Rodney on difficulty of detecting 1 ppm analytes with a FID - if you have significant separation between your peaks of interest and solvent components, 1 ppm is no problem with a splitless or on-column injection with FID detection. You are not in this situation. Even using my suggestion to get better early eluting peak shapes and lower split ratios, you still may be left unsatisfied with your sensitivity increases.

Valco makes a Pulsed Discharge Detector that would easily see 1 ppm in split mode. It costs several thousand and you need very clean bath gas to feed it. Argon is a good place to start; Valco will tell you what exactly you need.

Or, send a sample to your instrument manufacturer and have them test on an on-column injector if you can't afford to buy one and see for yourself. Instrument/detector manufacturers love to see if they can solve your problems by potentially selling you more equipment. An on-column injector will run several thousand too...

Or, if you really can afford it, buy an MSD and do selected-ion monitoring. You will be overwhelmed with sensitivity in SIM-mode!


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Rodney George on Friday, October 4, 2002 - 07:55 am:

Spencer

I won't say you are wrong BUT:

I would love to see any chromatograms of the analytes in question (EO and ethylene chlorhydrin) at 1ppm from water-methanol mix on a standard sentivity FID. I have reports of efforts several months in length trying but failing to achieve this (from a rather difficult matrix, I should add).

They can be seen, but reproducible and with a low RSD adequate for a FDA validation, that is another question.

On the positive side, you are correct that the Valco PDD will see these accurately at 1 ppm, and with the reproducibility desired, but you didn't hear it from me. :-)

with all best wishes,

Rodney George
Senior Research Chemist
Supelco Gas Separations Research


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Spencer on Friday, October 4, 2002 - 07:06 pm:

Brett,

Having not read these papers, I can't comment on their quality, exact methodology, or absolute detection limits. One is a rather dated, but may still point you in directions toward success. If you are doing this for FDA documentation, my guess is that you'd definitely want to invest in proper equipment if these references don't help you out.

Simple and accurate method for determination of ethylene chlorohydrin in dried spices and condiments. Aitkenhead, Paul; Vidnes, Arne. Norw. Food Control Auth., Oslo, Norway. J. - Assoc. Off. Anal. Chem. (1988), 71(4), 729-31.

Abstract

A simple and accurate method is described for the detn. of ethylene chlorohydrin (ECH) by using capillary gas chromatog. (GC) and flame ionization detection. Acetonitrile-methanol was chosen as the extn. solvent in preference to other solvents because its use reduced the no. of compds. detected by the GC system, thus enabling easier identification and quantitation of ECH. The coeff. of variation for the method is 2.7% at 5 ppm, and recovery is good, even for the std. addn. of 1 ppm. Fifteen different spices and condiments were analyzed using this method; 20% were identified as pos. for ECH. The method also identifies the related compd. ethylene bromohydrin (EBH).

Reference 2:


Gas chromatography of ethylene oxide and its toxic residues. Ben-Yehoshua, S.; Krinsky, P. Volcani Inst. Agr. Res., Bet Dagan, Israel. J. Gas Chromatogr. (1968), 6(6), 350-1.

Abstract

The toxic residues obtained in food products fumigated with ethylene oxide (I) were detd. qual. by gas chromatog. of Me2CO exts. The following 3 chromatographic systems were used: a Poropak-R column at 140°, with N carrier gas and an H flame ionization detector; a 10% polypropylene glycol/ Chromosorb W column at 65°, with N carrier gas and an H flame ionization detector; a silanized Gaschrom P column contg. 4% SE-30 +2% XE-60 at 110°, with He carrier gas and a thermal cond. detector. The 2 former systems permitted the detn. of <1 ppm. of each compd. The retention times relative to I on the Poropak-R column, retention times relative to I on the polypropylene glycol column, and retention times relative to I on the SE-30-XE-60 column are given: ethylene glycol, 30.0, -, 3; ethylene chlorohydrin, 24.0, 24.4, -, diethylene glycol, 52.0, -, 17; Me2CO, 3.3, 1.8, -.


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Brett Libby on Monday, October 7, 2002 - 03:09 pm:

First off, thanks for the help and advice so far. Since I'm new to GC all this info is helping me to learn alot.

As far as my specific issues are concerned, the simple story is this. I have an old method to detect Ethylene Oxide and Ethylene Chlorohydrin down to 1ppm using a packed column:
3% Carbowax 20M on Chromosorb 101 (80/100mesh) 6ft x 2mm ID glass column
Two microliter injections were used and FID detection was utilized.

Thus, I was hoping to find a capillary column to do the same task. Any ideas?


Top of pagePrevious messageNext messageBottom of pageLink to this message  By j de zeeuw on Tuesday, October 8, 2002 - 12:40 pm:

looking for traces is always difficult. 1ppm and 1 ul injection will give you 1ng on the column. The water/methanol mixture is also not most favorite as it evaporates slowly and take a lot of heat from the injector.
Sometimes less injection using a more selective column and high flowrates are better... Dow did some good work on that doing direct injection technique they called "micro column injection".If you need this paper, let me know.

An other serious thing to concern is that you need a highly inert and efficient column for eluting such trace compounds.

We have reports showing good performance at high levels but losses at low levels using this (low-end) brand supplied by Phenomenex. Deactivation is essential which the ZB brand does not seem to care about. Anyway, think of other suppliers that can help you by sharing technical expertise like J&W, Restek and Varian.
Their helpdesks will be happy to help you out on one of their high quality capillary columns offering a helping hand...


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Rodney George on Wednesday, October 9, 2002 - 05:20 am:

to Spencer and all

I realized yesterday that communication is not always complete and often certain assumptions are not always shared between posters and responders on e-lists such as this one.

For example, when the initiator of this thread spoke of a 1ppm methanol-water std or sample of EO or Ethylene chlorhydrin I "assumed" he/she was discussing a std 2% (20mg/mL) methanol-water solution of a compound that contained a 1 ppm impurity, as might be done for a USP-FDA analysis, not a 1ppm impurity in the solution of methanol-water itself.

There is quite a difference, 50X ! to be exact in magnitude.

I apologise for my misunderstanding if I did misunderstand the original problem and I ask posters to clearly present their problem parameters to the e-list.

Our assumptions are guided by our work experience and often do not correspond to the intentions of the originator of the problem.

Again, my apologies for any false assumptions I brought to the conversation. I will try to do better in the future.

Rodney George
Senior Research and Development Chemist
Gas Separations Research
Supelco
595 North Harrison Road
Bellefonte, PA 16823


Top of pagePrevious messageNext messageBottom of pageLink to this message  By venkitesh on Tuesday, November 5, 2002 - 03:53 am:

I could not find a better place to put in this query. I have to do an analysis by cool on column injection as per the standard method given to me. But I don't have a cool on column inlet, but have a split/splitless inlet. Can I do something about it and manage the injections.


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Anonymous on Tuesday, November 5, 2002 - 05:40 am:

If the compounds are thermally labile and break down in a hot injection port you will not be able to use a split/splitless injector. Try splitless mode at a relatively low injection port temperature and see what happens.


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