Challenging Application (DCPD/Hydrocarbons analysis)

[cawarhur]

I've been tasked with devising a GC protocol for a relatively challenging application. The samples we analyze come from the steam cracking of petroleum which are then typically subjected to some form(s) of treatment afterward, such as distillation or heat soaking. In the streams we look at, the major component is DCPD (dicyclopentadiene), which undergoes a retro Diels-Alder at 150C forming CPD (cyclopentadiene), and therein lies our problem.

Because we have temperature sensitive analytes, we hold our inlet at 160C (6890, Rtx-5 column, split mode) which is definitely not hot enough. However, after heat-soaking the material, we get trimer, tetramer, and possibly higher MW oligomers that all boil well beyond 160C. You can probably see where this is going...

Our problems:

1) Can't heat inlet too high because of temperature sensitive analytes, but need to vaporize high MW analytes--problems with discrimination and inlet contamination
2) Can't run the GC too hot or we see rearrangement products, but need to in order to cause elution of high MW analytes
3) Residual C5's in sample may be lost to backflashing

What I've done so far is affirmed that the GC method used to analyze these samples (set up before I began) is inadequate by running a variety of tests using trimer/tetramer standards, solvent blanks, column bakes, and inlet bakes. By various methods, I also confirmed that our contamination source is the inlet. With a slight pressure ramp and some tweaking, I've also come up with a method that elutes our higher boilers in a reasonable amount of time.

I think the obvious step is to try running in splitless mode to address a large portion of the difficulties we are having. I'm hoping that running in splitless mode with perhaps a double-tapered liner will be the silver bullet for this application as it will address backflashing, injector discrimination issues, and possibly even the inlet contamination we've been seeing (because of the residence time of the samplein the inlet, possinly vaporizing the high boilers).

My question to the community here is what are our other options?

[Karen01]

Maybe two methods, with the lower boilers being done by headspace if possible.

- Karen

[Don_Hilton]

For the inlet issues, either on column or a PTV inlet might work for you. You get the compunds on the column at a low temperature and then heat either the column or PTV inlet to get the other compunds moving once the lower boiling, temperature sensitve compunds are out of the way. Thus, you can start the compunds moving on a GC column at temperatures below 150 degrees and end with an inlet temperature above 250 (or more), if you want.

[chromatographer1]

You have identified your situation accurately, based on my experience with CPD.

I would suggest a megabore column, fairly low film thickness and using a valve injector might be an alternative. This avoids the discrimination issue. I would also include a short retention gap with no phase to improve vaporization of the higher molecular weight components. A faster flow rate will also improve the sample transfer out of the sample loop of the valve. I would suggest a 0.1 microliter as the minimum volume and 0.2 microliter as an optimum.

The heavier MW components will refocus on the initial portion of the phase coated column and will elute normally as you program the oven to higher temperatures.

Negative issues might be inadequate vaporization, contamination of the retention gap, and inadequate prewarming of sample through the valve.

But, you wanted options, so here is what I can suggest.

best wishes,

Rod

[AICMM]

cawarhur,

I would second Don's vote for a PTV type injector since you can temperature program. This will probably end up with a megabore column (most of them I have seen do.)

I would not advise splitless since residence time will be higher in the inlet. I also suspect you are talking pretty high concentrations which will flood splitless.

Best regards,

AICMM