The Impact of a Dirty Marijuana Extract on Pesticide Recoveries from a Splitless GC Inlet

In a previous post I told you about our experimentation with marijuana (that didn’t sound right…) and pesticide extraction via QuEChERS.  An important part of that work was cartridge solid phase extraction (cSPE) cleanup with 500mg CarboPrep 90 / 500 mg PSA cartridges.  To minimize losses of planar pesticides (e.g hexachlorobenzene), which we know can occur on carbon, as mentioned in another blog, we eluted the cartridge with acetone:toluene, a tip we picked up from Jon Wong and Frank Schenck (retired) from the US FDA.  To evaluate the success of this approach, we quantified spiked marijuana extracts that were QuEChERS extracted and either cleaned up or not cleaned up with cSPE.  Or should I say, we attempted to quantify the unclean extract?

Let me go back a bit and give you our Agilent 6890 Split/Splitless GC inlet conditions to set the stage for the recovery table I want you to review.  For a 1µL splitless injection we used a Restek Premium 4mm ID single taper inlet liner with wool at 250°C and a purge valve time of 60 sec.  The GC column flow was 2 mL/min, which gave a GC inlet flow of 1.6 mL/min.  Now some of you are probably saying, “Waitaminute, how did he know that inlet flow?”  It’s because I use a handy, dandy Pressure/Flow Calculator that you can download free from Agilent.  Others are wondering, why does the inlet flow matter?  Well, it’s because a general rule of thumb for a splitless injection purge valve time is to set it so the inlet has been swept approximately 1.5 to 2 times before opening the split valve.  This should ensure a quantitative transfer of analytes from GC inlet to GC column.  My sweep with the 0.9 mL volume single taper liner was 1.7 times.  Almost perfect, right?

However, if the sample extract is particularly dirty, and especially if it contains a high concentration of non-volatile material like chlorophyll, the analytes have to “chromatograph” out of the co-injected non-volatile material to get to the GC column.  And that may be more difficult if the liner contains wool, and it does, in this case.  This leads to seriously reduced recoveries (via poor transfer) of less volatile analytes of interest, which can be seen in the table below.  Hopefully you astute blog readers out there (do we have any other kind?!) have noticed another effect, and that is degradation of DDT, which inflates the DDD result, since DDD is a degradation compound of DDT.  Also note the complete loss of Dicofol in the unclean extract, a DDT-like pesticide that is even more sensitive to “dirt” in the inlet.  THESE RESULTS ARE FROM THE FIRST INJECTION OF THIS DIRTY EXTRACT!  YES, I AM SCREAMING AT YOU!  But only in print…

OK Jack, get to the point.  In summary, cleaning up your sample extracts for GC can be VERY important, especially for ruggedness of any method.

Cartridge SPE cleanup of a QuEChERS marijuana extract with 500 mg GCB / 500 mg PSA.

Pressure / Flow Calculator showing GC inlet flow, which helps estimate a proper purge valve time for a splitless injection for a given liner volume. In this case, a 4mm ID single taper liner with wool was used.

Pesticide recoveries for later eluting compounds are reduced in the unclean extract (No cSPE) due to poor transfer from the inlet liner to the GC column. In addition, DDT and Dicofol are degraded in the inlet due to injected “dirt” from the sample extract.

5 Responses to “The Impact of a Dirty Marijuana Extract on Pesticide Recoveries from a Splitless GC Inlet”

  1. Dave Tai says:

    Hi Jack,
    What is the ratio of aceton:toluene to elute Hexchlorobenzene? How much volume to elute? Is sample in acetontrile before cSPE ? Does the elute contain all target pesticide as well as Hexchlorobenzene? Can we use the same mix (aceton: toluene) on dSPE as well? Sorry about so many questions.



  2. Jack Cochran says:

    Dear Dave:

    We used 15 mL 3:1 acetone:toluene (vol:vol) to elute hexachlorobenzene from the 500 mg GCB / 500 mg PSA. This was an “optimized” amount based on looking at sub-fractions and monitoring the cartridge SPE elution profile of planar pesticides, including hexachlorobenzene and pentachlorothioanisole, among others. The eluate does contain most of the target pesticides, nicely recovered; however, in some cases, for reasons we’re not quite sure of yet, in the marijuana extracts we see low recoveries for thiabendazole, cyprodinil, and fenhexamid.

    The sample is indeed in acetonitrile (from the QuEChERS extraction) prior to being loaded on the SPE cartridge. We use 0.5 to 1 mL extract typically when doing these complex samples (e.g., marijuana, botanicals, herbals, spices, tobacco, tea, etc.) to avoid overload of the cartridge.
    Jon Wong of the US FDA did some nice work using toluene to recover planar pesticides during dispersive SPE cleanup with GCB. I believe he used 1 mL extract, shook it up, and then added a small amount of toluene to the dSPE tube to recover the planar pesticides, centrifuging, and taking the supernatant. Since he was using LC-MS/MS for determining pesticides, he evaporated the solvent off from this clean extract and re-dissolved in mobile phase. I did some similar work and added 0.75 mL QuEChERS extract to a 1 mL dSPE tube containing PSA and GCB, followed by adding 0.25 mL toluene, shook, centrifuged, and then analyzed using GC-MS. The GCB will remove chlorophyll (and some other matrix compounds), but the toluene helps keep recoveries of planar pesticides relatively high. A side benefit of doing this is when injecting a blend of acetonitrile:toluene , early eluting peak shapes are better than for acetonitrile alone when using a relatively non-polar pesticide column like Rxi-5Sil MS (

    I hope this information helps, and if you pursue any of the work, I’d be interested in your observations.



  3. Dave Tai says:

    Hi Jack,
    We tested 3:1 aceton: toluene as elute and it works for all pesticide targets we are looking for but with one side effect. Since SPE packed in plastic syringe , toluene seems attacking plastic and left with large amount “PE” in the extract which I could see on FID and also identify it on MS. I am not sure it was coming from syringe or the frit disk that hold material in place.


  4. Jack Cochran says:

    Hi Dave:

    Interesting observation. I don’t remember specifically noting that in our work. The frit is the most likely culprit, due to its surface area and being directly in the flow path. Can you tell me what m/z ions you were seeing on MS as you observed the contaminant (or was it multiple contaminants)? That way I can take a look at some of my old data, too?

    Thanks Dave!


Leave a Reply