Purge & Trap linearity problems

[ArjanH]

Hello everybody,

I am facing lots of diffeculties setting up my new instrument. The problem I see is that the response of the aromatic compounds in my system is not linear as the response of the chloronated componunds is linear from 0.2 ppb to 400 ppb. The aromatic components show a curve where the response in the range from 0.2 ppb tot 50 ppb is way to low for the calibration curve. The calibration curve for the aromatic components shows more a exponential curve.

The method is comparable to the EPA624 and my system settings are:

EST Encon evolution Purge & Trap with EST Centurion autosampler and a Vocarb 3000 Trap equiped with moisture reduction trap. Purge- and desorb settings are taken as sugested by the supplier.
Sample loop and purgevessel are 5.0 ml.


The GC configuration:

7890 GC
5975 MSD
split-ratio 100:1 (on advice we also tried splitratio of 40:1, without succes)
Column: Agilent DB624 30m 250 um ID, 1.4 um film.
temperature program: the problem seems independent from the program. We tried different programs but all programs show the trend.

Any idea's on the cause of this problem?

[rb6banjo]

I assume that your aromatics are BTEX? What are the identities of your chlorinated analytes? Do you run an internal standard(s)? How's the precision on that material? How are the standards made - by the autosampler or by you? I don't know the EST Encon. We have Tekmar equipment in our department.

[ArjanH]

Yes, the aroamtics are the BTEX, the chlorinated components run from dichloromethane up to 1122 tetrachloroethane. Up to this point we didn't use an internal standard for the calculation. We also have a Tekmar Velocity XPT unit, this is the old instrument we want to replace and from historical reasons we never used an internal standard for that instrument as well. My guess would be that there is something with the trap or my standard preparation. I diluted the standards myself and filled the autosamplervials directly after preparation.

[rb6banjo]

I agree with you. It doesn't make sense that it's linear for the chlorinated aromatics but not the BTEX components. It has to be something in how you're preparing the standards. Can you separate them - run a BTEX only standard - and see what happens?

I get beautiful linearity for BTEX and trimethylbenzene with our Tekmar system from 0 to 50 ppbv - no internal standard.

[msidler1]

Arjan,

Have you tried a different VOCARB trap? If the trap was not made properly (i.e. perhaps there's a small amount of CarboPak X at the top, which would trap aromatics on their way off the trap but allow most chlorinated compounds to pass on by), that could explain your issues. The only other explanation I can think of is the Standard mix. Are all of your target compounds in one mix, or do you have separate mixes? Is it possible that one of those mixes contains only the aromatic compounds, and perhaps that mix may be bad?

If you'd like to discuss your issues further, I work for EST in tech support so feel free to give us a call at 888-890-1404 and choose to speak to a tech support specialist. I, or one of my colleagues, would be glad to assist.

Hopefully we can figure out what's going on and get you back up and running. Let me know what you find out. If you'd like to PM me or send me an email, please do -- my email address should be in my profile.

Mike

[Peter Apps]

Am I right in understanding that if you fit a straight line through the BTEX calibration the lower points are below the line ?

This does not necessarily mean that the response at the low end is too low. It could just as well be that the response for the higher levels is too low.

Low response at the high end might be due to going above the linear range of the detector - do you have similar quantities for the two classes of compound ? If you look carefully at the results with a lower split ratio, does the line curve at a lower level than it does for a higher split ratio ?

Peter

[Bigbear]

I suspect water in the source. Try cleanning the source and repeating your calibration. Or run your calibration high to low on the "dirty" system. If things get better, look for an active site.
There is a known issue with the Agilent 73 and 75 MSD's dealling with water /purge and trap.
If you can swing it you would be better served to use a 20M X 0.18 column. Lower column flow= less water.

[James_Ball]

Purge and trap can be a very confounding system. I have been running them for over 20 years and one little change can make you go from having every compound running as average response factor calibration with less than 5% RSD to having half of your compounds running that way and the other half on quadratic fits that look like ski slopes.

If the early analytes have wider peaks than the later ones, check the glass wool packing at the ends of the traps. If there is too much packed too tightly it will cause that problem. I know from experience as I have had it happen when I pack my own traps similar to the VOCarb3000.

From the sound of your aromatic curves it sounds like you have something in the system that is absorbing specifically the aromatic compounds, once you have those sites saturated a larger percentage of your analytes will pass through which causes the exponential rise in peak areas with increasing concentration. Think of it as up to 50ppb the contaminate is absorbing most of your analyte, then at 100ppb instead of maybe 50% passing through you get 65%, then at 200ppb you get 80%, then at 400ppb you get 95%.

If you have ever foamed a dirty sample it can deposit residue in the system that can absorb certain classes of analytes more than others. Is this a pristine purge and trap system or one that has seen dirty samples?

Cold spots can also cause linearity problems, make sure where your transfer line is attached to the injection port does not have too much line exposed to the cool laboratory air, I always wrap a little glass wool around the joint up to the injection port.

If you have an IR thermometer check the trap temperature during desorb and bake to make sure it isn't over heating or under heating. Under heating it may not desorb all the compounds equally, over heating can break down the packing material which can cook off and cause system contamination and adsorption problems.

Another thing to try with the Agilent systems is to use the larger Draw Out Plate in the source, it helps with linearity. Also we have found that the ETP film type multipliers are not as susceptible to gaining response at higher concentrations of standards than the older K&M style.(not sure about the new Triple Axis detectors, I haven't run those yet)

Also try to use the same microliter syringe to make each level of the calibration. I have found if you use a 10ul syringe for the low standards, 50ul for the mid standards and 100ul for the upper standards it can cause problems since they may not be exactly true to each other volume wise. Even if you have to make multiple draws to spike the standard, it seems to be more accurate. I tend to use different size volumetric flasks, larger for the low concentration standards even use a 1L at times, down to a 50ml for the high concentrations so I can use only one syringe to spike with.

Errors can be in prep, purge, trap, heating, contamination, GC inlet, GC column, or in the Mass Spec. Volatiles has so many places for errors it makes trouble shooting way more difficult.