Loss of Hydrocarbon Sensitivity on FID

[la_donz27]

Hi,

We’re setting up two new 7890 GCs for a gas analysis of C1-15 using a DB-1 30m x 0.32 mm ID X 1um film. Standard is introduced using a GSV and sample loop into a Split Splitless injector. It’s a split injection (20:1) with H2 column flow of 3ml/min, temp ramp of 40C up to 230C over 18 minutes, with FID detection. We have eight other Agilent 7890s set up with this same method, and we use the same heated gas standard on all ten units. Sample loop sizes are all the same, split ratios same.

The units and consumables are all brand new out of the box.

My new units both have their methods set up exactly as the eight working units. We are using Hydrogen as makeup gas instead of Helium on one unit, and Helium on the other. This doesn't seem to be the problem as they are both responding in the same manner:

Both new units have virtually no sensitivity for any compounds after toluene. I don’t see ethylbenzene, xylenes, etc, up to butylbenzene. However, on the eight working units, those peaks show plenty of sensitivity (>25 000 AC).

We did inlet pressure tests (seem ok, pressure loss over 10 min <1psi), baked and trimmed the column, re-installed the column a second time, changed the liner, tightened the injector weldment nut, and cleaned and reassembled the FID.

I’m at a loss as to why these two new GCs have such poor sensitivity (on an FID no less!), and what's worse, we’re on a timetable to get them validated.

Anybody have any suggestions where else to look? What are we missing?

[Peter Apps]

If the instruments are new then this is Agilent's problem. Discrimination on molecular weight sounds very much like a cold spot to me.

Why are you using hydrogen as make-up gas ? - the only effect is to double the fuel flow to the flame. The best makeup gas is nitrogen, you need something heavy to slow down the diffusion of hydrogen against the flow of gas in the flame.

Peter

[dblux_]

Turned off or defective inlet heating (?)

[LabProARW]

If no peaks of any sort elute after Toluene, does the baseline go flat like the flame went out right after the Toluene? Are their any hidden events/settings set to change after Toluene like baseline compensation set by accident? Extending the run does not show the lost peaks at later times? Sample of a different matrix do the same strange thing?

[la_donz27]

If no peaks of any sort elute after Toluene, does the baseline go flat like the flame went out right after the Toluene? Are their any hidden events/settings set to change after Toluene like baseline compensation set by accident? Extending the run does not show the lost peaks at later times? Sample of a different matrix do the same strange thing?

Thanks for the good ideas.

Here's an update - we've gotten the peaks after toluene back and also they have gone from short, broad and tailing to sharp and not tailing. This was achieved by lots of little tweaks with the flow, split, and hardware, but the most interesting being turning off the Gas Saver. Doing this is where we saw the tailing disappear. Why? Can't answer that, as the eight lab GCs have the Gas Saver on. But we noticed that the missing or ugly peaks started when the Gas Saver went on at nine minutes.

We think there's something wrong with the inlet and it's under warranty. We're going to swap it out and see what happens.

However, we still have one anomaly with the analysis. Remember that I said that all ten GCs are configured exactly the same? On the eight lab GCs, the last four peaks of the standard elute A,B,C,D. On the two new GCs, they elute A,C,B,D.

That one's a mystery we haven't yet solved. I thought we should put one of the two offending columns into one of the eight lab GCs to see whether the phase has some manufacturing difference, but currently the powers that be don't want to try that.

[la_donz27]

The first run on the new split/splitless injector is now perfect. Including the elution order of the last four compounds.

Go figure - what are the chances that a brand new 7890 will have a wonky inlet? Oh well, all's well that ends well. And within one day of our ship deadline!! Whew!

Thanks again for all the ideas.