Chromatography Forum

Hello all!

I have been reviewing my most recent calibration for EPA 8260 and I am noticing a major rise in internal standard area as my calibration standard concentrations increase. I am seeing area recoveries of nearly 60-80% higher for the highest points vs blank and lower cal points. I recently switched from a Tekmar 9 Trap to a Vocarb 3000 in order to reduce water intake to the inlet. Using the 9, I didn't see as much drift with the internal standards (+/- 10%), but using the Vocarb, I am seeing substantial fluctuation. Lower concentration injections appear to align with their point on the calibration. If anyone has any insights, it would be greatly appreciated. I am currently running an agilent 7890B and 5977A with a Tekmar Atomx autosampler. Also my split ratio is 100:1.

Thank you!

Do the area counts rise with increase in target analytes or is it with and increase in methanol content?
You can test this by adding extra methanol to low level standards.
I see an increase (50%+) in 1,4-dichlorobenzene-d4 after I have analyzed samples containing NAPL (creosote or such). Curiously it doesn't affect the 1,2-dichlorobenzene-d4. The clue is that the increase is related to the amount of added methanol, if I add 100uL of methanol to all standards the areas are constant.

The cure is to flush the transfer line and inlet to remove the leftovers in the system, along with a new trap.
I'm using a OI Eclipse and Archon with a Vocarb trap.

This is a common problem with purge and trap analysis, and seems to occur across different systems, not just isolated to one manufacturer.

From all the testing I have done over the years it is often more related to compound concentrations than just methanol volume. When I have made standards from different concentration stocks so that the methanol volume is the same I have still had the problem occur. Even on a system that has only seen drinking water samples.

I have not proven it yet, but I believe it happens when you get some carbon buildup in the system, possibly from the heat of baking the traps or transfer lines causing charing of residual hydrocarbons, which then act as an adsorbent coating. It affects the higher molecular weight compounds more than the lighter ones it seems which would be in line with what a charcoal trap would do. It acts similar to a chromatography column where the gasses and lighter compounds make it through while the heavier ones don't. As you load up the coating with higher levels of compounds, it saturates the sites that can capture the analytes thus allowing higher amounts to pass through to the GC when you have higher standards. When it gets really bad I have seen all of my last few eluting analytes give highly quadratic calibration curves.

I had one Archon/Encon system that had this problem really bad after running samples that were very high in TOC content(PVC momomer waste samples). No amount of flushing would completely cure the problem, then we had the system refurbished with all new plumbing installed and the problem went away and we had perfectly linear calibration curves across the board with less than 10% drift of the 1,4-Dichlorobenzene-d4 internal standards. This lasted until we had to run those samples again about six months later. I think the major problem for us was cause from the anti foam agent we had to use with those samples, it contains a heavy weight hydrocarbon of some kind I think, maybe octanol or some other very heavy alcohol.

One thing that seems to help prevent it is to run the moisture trap back at 180C instead of 260C. I am thinking lower temperature leads to less chance of char formation, but that is just a guess.

I think James is partially right. I have seen the same thing on traps. However, if you sit down and do the calculations (a real pain, but educational) you'll see that the purging process itself is concentration dependent. Good ol' Henry and Raoult - there are days I'd like to shoot both of them!

Anyway, as the total concentration of the analytes in solution rises, so does your purging efficiency. If you go back and read really really old USEPA docs (I'm dating myself here) you'll see that they struggled with this, and they weren't working anywhere nearly as low as what we can all routinely do today.