Non-Linear FID response to C18:0 FAMEs

[II5tickmanII]

Hi and thanks for the advice all. The Young Lin 6000 series are very similar to HP5890/6890s (i had never seen one before either but looking at them they are esentially the HP GCs in a different case). I too expected the oppposite effect by dropping the inj temp however through troubleshooting and with the limits of my knowledge dropping the inj temp to that of a previous method was the only option i had left. Routine samples and p-test material give typical / expected results (in terms of c-chain %s of total) however i am keen to get to the underlying cause whatever it is.

The method i am following is current ISO practice (5508:1995 and 12966-2:2011) with a 1:15 dilution stage at the end

Peter:
Yes, I am still experencing discrimination against the heavier c-chains
My injector settings are
Sample Volume: 1.0ul
Pre-Inject Delay: 0s
Post-Injec Dwell: 1s
Injection Speed: "1" an arbitrary value given by the clarity software that i can only describe as not fast... increasing number = increasing speed, up to a value of 5 i think
There are 5 solvent flushes prep and post injection of 3ul Iso-octane

BigBear / Krickoss:
The solvent expansion in the liner should not exceed that of the liner volume. I have checked this using the calculation on the Restek site (desilanised Restek precision liners used with glass wool insert)

Alexandre:
There is no hump/baseline noise increase with inj temp increase, increase in oven temp obviously does give an increase in baseline but no humps appear which may be masking peaks

Once again thanks for the help, its much appreciated.

[Peter Apps]

Try setting the post injection hold to zero, and make the injection speed faster.

Peter

[II5tickmanII]

Try setting the post injection hold to zero, and make the injection speed faster.

Peter

Been a while but finally got a chance to perform some trials with faster injection speeds. Worked a treat!

Much more linear area responses across concentration ranges from 0.5 to 4.0mg/ml and also managed to get the Calculated Response Factors reasonably close to those of the Theoretical Response Factors.

Thanks for all the help!

[Peter Apps]

Great, thanks for the feedback.

Peter

[jdezeeuw]

Interesting discussion. I had a similar one recently.

To high injection port together with no wool may give strange effects.

Please make sure you optimized injection: for FAME use Glasswool and make sure its positioned in the top of the liner, so you inject into the wool; Use 2mm ID liners. If possible use hot-needle injection (not always possible with Agilent systems); Inject smallest possible amount. It will help you in evaporation and keeping inlet clean. As long as you get your detection limits of course.

NEVER use pressure pulse with a splitted injection please. this is a good way to get strange results. This pulse was developed for splitless injection.

also I hope we look at peak area and not peak height.

jaap\
Restek Corporation

[Peter Apps]

Hi Jaap

From the symptoms in this case I think that the problem was caused by fractional distillation from a hot needle causing discrimination against high boilers. With the pre-injection dwell eliminated (and so a cooler needle) the problem was solved.

Peter

[jdezeeuw]

This good for an interesting discussion theme for one of our meetings/troubleshooting sessions: How high should my injection port temperature really be? I think we can add a lot of variables here making the best recommendation a real challenge. If temperature/mass is that important we will have to deal with"What works in one system does not have to work in another system.."..

[Peter Apps]

Agreed.

And a temperature that is plenty high enough with a brand new inlet liner will be too low once a bit of crud has built up.

I still maintain that due to historical inertia current inlets are simply the top 3 inches of a packed column with the packing taken out, and that if someone was to sit down and design a capillary inlet from scratch it might not look anything like what we now have.

Peter

[II5tickmanII]

Apologies for the lack of response Peter and Jaap. I have only just seen this discussion. As a person who has attended and benefits from troubleshooting seminars i would suggest this would be a good topic.

Peak area is what i have used in each measurement. Also for information (you'll be pleased to hear this Jaap) it is a Restek SKY Precision liner #RE23305.1 and an Rtx-2330 column 30M x 0.32mm x 0.20um. I then make injections of 1.0ul with a 30:1 split. Inj and Det both at 250C

The oven programme goes in stages from 100C to 190C then a burn off at 225C and allows for peaks C4 to C24:1 to be read within a 20min program including a small burn off at the end.

I am now looking to re-calibrate the instrument using these preferred settings, which will hopefully allow for generation of more accurate results.

In your opinion, how many injections do you think a column as above could be subjected to and still remain effective? We have approx 700 sampes go through the column each month and I was hoping that i would get a few months out of each column, there has been no real impact on efficiency, resolution of major peaks and RT but i have noticed that tailing is now becoming more severe and peak asymmetry is not good. This is after approx 1700 injections.

Kind Regards

[jdezeeuw]

Hi
Lifetime of GC columns is a good topic which we discuss in most meetings. The bonded phases were developed to increase lifetime as well as the surface bonded phases. This eliminated the risk for droplet formation. It only takes a few droplets to ruin efficiency of columns. As this problem mainly occurs with splitless injections, split sampling is more forgiving (no condensation effects). The 2330 phase is a NON-bonded phase which is per definition more vonorable. Most of the trouble will be activity/contamination which will happen in the first section
So the solution for longer life time:
- inject less (but enough to see your lowest levels);
- do more sample cleaning (of course,we don't want that);
- cut after XX analysis 20 cm of the inlet ( peaks will slowly move)
- connect with a guard column ( this can be a defined length and this lengthy is replaced every YY injections. You have constant retention times;
- if you have aggressive sample use a coated guard. I always say:: buy 2 similar columns and use one of these for taking sections of 2 m and use that as guard. As its coated it will take much more sample stress.

I am sure there will be more smart solutions for extending life time..
jaap
restek corporation