FAME analysis inlet activation?

[sdegrace]

Thanks very much for your thoughts, Rod and Bruce! You guys have given me a lot to focus my mind and point me in hopefully the right direction. I'm going to mull this over during the weekend and attack the problem fresh on Monday. When (if, heh) I figure this out I'll post and let you know what I came up with!

Stephen

[Tony]

Hi,
I woudl agree with the previous comemnts on discrimination. Is your customer also using a 6890? If so, try and replicate there injection conditions (needle depth into liner, injection speed, number of washes). The needle should go into the center of the glass wool region. We routinely do quantitative FAME analysis and find that hardware effects on Rf values are ofter related to bp discrimination and can be best overcome by optimizing injector parameters. We find that the Rf of even very different FAMEs, like 16:0 and 20:5, should be 1.00 under optimized conditions (using 17:0 or 23:0 as a RF reference).


cheers
tony

[chromatographer1]

I agree that inappropriate numbers for response factors is most often caused by injector discrimination.

The second most likely cause is the loss of unsaturated FAMEs due to degredation from the reaction to esterify the FAs.

The third most likely cause is loss in the injector from chemical reaction-adsorption from active sites in the injector, including the metal needle left in the injector during the vaporization of the sample.

The players are cast. The dialog is written. The performance has been witnessed. Now the critical review is waiting to be written. I can hardly wait to hear the end of the story.

best wishes,

Rod

[sdegrace]

I'm sorry I haven't replied in a week, I got drawn into a completely different crisis and I haven't been able to get back to this problem until now, as usual just when I need to have the problem solved lol.

I tried taking apart the inlet, swabbing everything with IPA, replacing the septum and the liner (used an SGE FocusLiner from a foil-wrapped package), removed the fitting with the gold seal and washer, sonicated that fitting in IPA, replaced the gold seal and washer, and did my run - so I've made a good faith effort to test this theory I think.

Right now, the calibration standard, which is home-made from Nu-Chek reference materials, has perfectly conforming RF values which is great. Unfortunately the Nu-Chek 21-FAME standard is again way too high (RF=~1.15). I don't honestly think there is anything wrong with the N-Chek oil per se, which leads me to think that maybe I did something to it when I diluted it with iso-octane + 0.05 g/L BHT.

I'll re-read over your comments and keep you posted as I try and fix this!

[chromatographer1]

Don't forget to run a blank in case your isooctane or BHT has a peak right where you don't want an extra peak.

best wishes,

Rod

[sdegrace]

My blanks are clean, I'm seeing only iso-octane and BHT. I'm using the same batch of diluent as was used to prep the standard as the blank.

I'll see what my overnight run with the new standard prep has wrought soon - fingers crossed!

[sdegrace]

I followed the cleaning procedure recommended by Agilent at:

http://www.chem.agilent.com/Library/sup ... a16022.pdf

I used a 32 caliber gun brush as there weren't any 38s at Canadian Tire... I'm glad I went with the smaller size, though, I can't imagine trying to stuff a 38 down there. Anyway, afterwards my RFs for the 21-FAME NuChek and the in-house calibration standard are all in spec, although the area percents in the Nu-Chek standard are all almost 1% higher than expected for the two unsaturated FAMEs and one saturated FAME of interest. *sigh*... back to ripping more stuff apart today!

[chromatographer1]

Thanks for keeping us updated.

Factor #3 from my list seems to have been important. Be aware: Certain decomposition peaks can elute under other peaks. You may have to perform separate calibrations with pure standards to be certain this is not happening.

Glad to see your diligence and I hope it all gets solved soon.

best wishes,

Rod

[sdegrace]

My problem with the percents just turned out to be something stupid with the integration parameters... when I had all the peaks integrated the area %s were good but with one of the unsaturated FAMEs of interest being low. Still in spec, but borderline enough that we've done further maintenance tasks to see if we can improve it. We've gun brushed the inlet again, replaced the septum, liner and gold seal again, replaced the split vent line, replaced the split vent filter, and we've taken apart the FID, replaced the insulating disks and the jet, and verified that everything looked clean, just for good measure.

I'll let you know if it looks any better after this treatment. If I have to order in any new standards to check for things, that'll be my next move.

Stephen

[sdegrace]

Just to update, we've found that we can improve the RF numbers through very rigourous maintenance, but we're convinced that the acceptable windows we've been given are unnecessarily stringent. We now have a good body of data to support the validity of results we've gotten to date with "bad" sys suit. We're proposing system suitability specs which we feel are reflective of a genuine measure of the suitability of the system for the intended purpose and are validating on that basis - so all's well that ends well! And I wound up learning a LOT about the HP split/splitless inlets .

[chromatographer1]

Thanks for sharing your experience. Many will profit (if they search the archives that is) from your generosity of time and information.

best wishes,

Rod

[Peter Apps]

I'm coming into this rather late, and it could be that you have already solved the problem. A usual I agree with Rod that the most likely source of this kind of problem is with the sample prep, but you are right in that this should not affect the mixed standard that you just dilute and shoot.

It is just conceivable that there is a molecular weight discrimination that is sensitive to carrier gas - hydrogen and helium have different heat capacities, conductivities and diffusion parameters. An inlet temperature of 250 C is maybe a bit on the low side for a C23 ester. It might be worth running an experiment with a higher inlet temperature, and if you cannot switch to hydrogen maybe you could get the other lab to run a set of samples with helium carrier.

Seeing that the exhorbitant price of helium is making labs switch to hydrogen - with plenty of threads on this forum - you might have stumbled on something significant here.

Please feed back to the forum if you do come up with anything.

Peter

[sdegrace]

Thanks for the inlet temperature suggestion. The next time I can run a sample where I have some leeway to play with the method again I'll have a look and see what happens (and let you know)! As to the carrier gas, the other lab can run helium, although they have since switched to hydrogen, and when I talk to them again I can ask if they have any historical data to show what their RF values are like with helium carrier.

As to carrier, we are aware that due to reasons of price and absolute world supply, in the not too distant future we are going to have to switch to hydrogen (and maybe have nitrogen carrier available as well if we find that any analyte we study is not compatible with hydrogen). We have already switched from helium to nitrogen for FID makeup gas, creating a substantial savings (I know of at least one lab that switched to hydrogen purely for throughput and still uses helium for makeup gas - ouchie!). However, from a GMP standpoint we see the change as a not trivial undertaking and it is something we want to plan thoroughly - we don't feel that this time that we're ready to go down that path just yet and we want to have a thorough plan in place. Unfortunately we don't have an instrument available that we can afford to switch to hydrogen and effectively take out of service for R&D purposes

Stephen