Chromatography Forum

Dear Chromatographers.

Due to the changes in the environmental legislation in my country we need to get a limit of quantification of 0.000001 mg/L for DDT. Our instrumentation is a PTV-GC/MS. We inject up to 300uL sample extract but it is quite difficult to distinguish the analyte from background even for clean water spiked samples, that´s why we don´t know if the analytes are extracted quantitatively. Next we will purify the extracts by SPE ( envicarb). Please, any suggestion?

Best regards.

Dear Chromatographers.

Due to the changes in the environmental legislation in my country we need to get a limit of quantification of 0.000001 mg/L for DDT.

This hurts my eyes a little: its 1 ng/L = 1 ppt (right?)

Our instrumentation is a PTV-GC/MS. We inject up to 300uL sample extract but it is quite difficult to distinguish the analyte from background even for clean water spiked samples, that´s why we don´t know if the analytes are extracted quantitatively. Next we will purify the extracts by SPE ( envicarb). Please, any suggestion?

Best regards.

We need some more information. What are the samples? What is your sample prep right now?

Having a PTV for large volume injection helps a lot in reaching low LOQs in GCMS. 300µL injection is a high amount and you need to be careful in avoiding backflash. Are you sure the solvent vent parameters are optimized for your inlet/solvent/liner? How do your calibration standards look like?

For that low of a detection limit you will most likely have to use SIM and even then it will be difficult. You may also have to concentrate to less than 1ml on the extract but if you are injecting 300ul you can't go much lower than 1ml. Are you extracting at least 1L of initial volume?

I was assuming that you ran in SIM. I agree with James, I wouldn't even try to reach 1 ppt in full scan. The more you concentrate your sample, the more interferences. Maybe if it's really clean drinking water, but we don't know which kind of samples you're trying to run.

Ok.
1) Yes, it is 1 ng/L. A pain in the neck.
2)The samples are tap or drinking water for this initial stages, after that they will be surface water samples. The extraction method is the classical separatory funnel liquid-liquid methylene chloride extration and K-D concentration. The sample volume is 1000mL cc to 2mL.
3) the PTV injetion port config. is: split ratio 25:1 ; 40° X 1min; 50°/min to 100° then 100°C/min to 250°. Column flow 1ml/min, const linear veloc. the liner is packed with deactivated glass wool ( 1.5cm for a 9.5cm liner)
4)The instrument is a Shimadzu-Ultra running in SIM mode, for DDT the ions are 235/237/165. We inject 120pg ( three portions of 100uL) of a std in pure solvent and the S/N for the 235 ion is about 200. the chromatogram looks fine in pure solvent std but in real spiked "clean" samples doesn´t. Least sensitive analytes Endosulfan-s and Endrin were lost, don´t know why.
5) The oven program is: 45° X 2min ; 20°/min to 300°. 30mX0.25X0.25um.

Thanks.

Why are you doing a split injection ? - you need to be closing the split when the inlet heats up to transfer the DDT to the column.

Peter

I understand the split at injection to vent the methylene chloride, but as Peter said, you need to go splitless once you begin to ramp the temperature, then hold splitless for about 1 to 2 minutes before going split again with a 50:1 ratio to vent off any high boiling compounds that bake out of the liner.

The fact that you have good results with standards would show you have enough sensitivity even if you remain in split mode. Are you using sodium sulfate to dry the extracts before solvent reduction? If you process blank water through the extraction then spike with DDT to make a standard from the blank extract do you still have the interferences?

If you are using sodium sulfate to dry the extracts, it could be there are contaminates in it. We have seen them before and either have to heat it in a muffle furnace or do a Soxhlet extraction on it prior to using it as a drying agent.

...Yes, we do splitless mode, we have observed the huge drop in the high vacuum level wich is a "indicator" of the DCM transfer process, it starts about the minute and some seconds after that go splitless until the 4 minute, then a split of 20 is set.
Also we performed the experiment of spiking a concentrated extract and the high backgroud persists, some analytes´ ions are distorted and noisy.
So far we are performing a scan run ( over a concentrated extract) to figure out why ( solvent, glassware) Things seem to be right only when we inyect a standard in pure solvent, any other process adds interferences.

That´s right, sodium sulfate could be an issue. We will soxhlet extract it, concentrate and analyze.

Best regards.

From your description you are not using the PTV in the way you should - leaving aside that you cannot decide whether you are running split or splitless.

While you inject, and for a short time afterwards you need a split ratio of about 20 or 30:1 with the inlet cool to evaporate off the solvent. When the solvent has evaporated you close the split and heat the inlet. Keep the split closed for 5 min or so (possibly longer) then open it to 50:1 to purge heavy crud.

You certainly should not be putting so much solvent down the column that you affect the MS vacuum.

Peter

Doesn't your instrument software have an inlet setting called "solvent vent" or "large volume injection (LVI)"?

It does exactly what Peter and James pointed out: split during and shortly after injection (evaporation of solvent), then splitless heating to get everything on column, then split again once everything is on the column.

Like i said in my first post, the parameters of this process should be carefully chosen/optimized or you risk backflash or losing analytes.

A "huge drop" in vacuum => this tells me that too much solvent is injected. So depending on your liner volume, you'll also have backflash.

Doesn't your instrument software have an inlet setting called "solvent vent" or "large volume injection (LVI)"?

It does exactly what Peter and James pointed out: split during and shortly after injection (evaporation of solvent), then splitless heating to get everything on column, then split again once everything is on the column.

Like i said in my first post, the parameters of this process should be carefully chosen/optimized or you risk backflash or losing analytes.

A "huge drop" in vacuum => this tells me that too much solvent is injected. So depending on your liner volume, you'll also have backflash.

A drop in vacuum can be expected, I even see it when doing a 1ul splitless injection with DCM, but also when doing cold splitless or solvent vent since part of the process is to transfer a little solvent to a cold retention gap to get a proper solvent effect for the lighter compounds. You hold the filaments off until after the solvent front passes through the system. If timed correctly you will be back to normal operating vacuum just before the first analytes elute.

The fact that the interference occurs only with extracted samples or spiked extracted samples and not standards spiked into pure solvent says there is something in the extraction process that is causing the problem here. Either excess moisture or a mixture of solvents, or an extractable interference is present. Hopefully the full scan run will help to show what the interference could be.

A drop in vacuum can be expected, I even see it when doing a 1ul splitless injection with DCM, but also when doing cold splitless or solvent vent since part of the process is to transfer a little solvent to a cold retention gap to get a proper solvent effect for the lighter compounds.

A blip in vacuum is entirely normal, but the OP says "we have observed the huge drop in the high vacuum level" which does not sound normal to me.

Nothing is anywhere near clean at the ppt level - hence the need to use the selectivity of SIM to pull the analyte signal out of the chemical background.

Peter

Endrin and endosulfan sulphate are easy to lose in the inlet due to reactivity with your liner/glass wool. Your large volume injection makes this more of a problem.

A drop in vacuum can be expected, I even see it when doing a 1ul splitless injection with DCM, but also when doing cold splitless or solvent vent since part of the process is to transfer a little solvent to a cold retention gap to get a proper solvent effect for the lighter compounds.

A blip in vacuum is entirely normal, but the OP says "we have observed the huge drop in the high vacuum level" which does not sound normal to me.

Nothing is anywhere near clean at the ppt level - hence the need to use the selectivity of SIM to pull the analyte signal out of the chemical background.

Peter

Mine will change from 7*10^-7torr to 5*10^-5torr for about a minute when I am doing 10-15ul injections. Not sure if this is what the OP would consider a huge drop in vacuum or not, but it is normal if using a solvent effect techinque.

Endrin and endosulfan sulphate are easy to lose in the inlet due to reactivity with your liner/glass wool. Your large volume injection makes this more of a problem.

I don't agree. Correct me if i'm wrong, but i was taught that a PTV process is better for these kind of molecules. Don't you have a higher chance of reactivity with hot injection compared to cold injection followed by heating?