Poor Linearity - Chloroform

[Jana P]

We are trying to develop a method for detection of acetone, ethyl acetate and chloroform in a new API.

We are experiencing issues with obtaining linearity of chloroform, even over a very small range (i.e. 80% to 120% of 60 ppm) which for our sample preparation is a concentration range of 1.5 µg/mL to 2.3 µg/mL.

Our instrument is Agilent 7890B GC-FID with 7697A headspace sampler. We are using a DB-624 column, 60 m x 0.53 mm x 3 μm film thickness.

All samples are prepared in 100% DMSO.

Headspace parameters are:

Oven 90°C
Loop 100°C
Transfer Line 105°C
Vial Equilibration 30 minutes
Pressurize Time 0.2 minutes
Inject Time 0.5 minute

With GC oven program being:

Oven at 40°C hold 10 min -> Ramp 100°C/min to 200°C hold 15 min.
Inlet at 220°C with split ratio of 1:1.
Colum at constant flow of 5.0 mL/min
Detector at 220°C

When we run our linearity solutions for Chloroform our peaks are all over the place. It is also not reproducible i.e. not always the same solutions which have the unexpected peak areas. For example one analysis the solutions at 100% look ok but at lower levels are bad but then the next analysis the solutions at 100% do not look ok but lower levels look ok. Note: performed with freshly prepared solutions.

We are not sure what is going on. When we analyse a standard and samples, we obtain reproducible results with Peak Area RSDs of 0.0%.

Does anyone have any advice on what could be happening? We have noticed that we also do not have an increase in peak area when we use a higher level of sample.

At the higher concentrations is my sample not getting onto the column because my inlet split is not correct?

Is there interactions between DMSO and Chloroform? Would I be best to try using DMF?

We have not looked closely at the other solvents yet as the linear range required for acetone and ethyl acetate is significantly higher than for chloroform

TIA.

[James_Ball]

With such high concentrations in ug/ml, I would increase the split ratio, at least to 10:1 you should still be able to see the lower concentration and not overload the column. Higher total flows will help with transfer of the analytes from the loop to the column and sharpen peaks as they get into the column in a tighter band.

[Jana P]

Thanks for the response James.
I increased the split ratio it was 1:1 (Split flow =5.0 mL/min) but I increased this to 1:6 (Split flow = 42 mL/min) to get the flow up at the inlet hoping this might help. It did not help. It does make the chloroform peak a lot smaller but reproducibility is still not an issue.
I still seem to be losing Chloroform somewhere.
I have two other analytes that i have no issue detecting but the chloroform is being lost somewhere.

Do you have any other suggestions?

Thank you

[James_Ball]

Have you ran Ethyl Acetate and Chloroform in separate standards? I am noticing that with a slow oven ramp those two analytes elute very close to each other, with the 100C per minute ramp they could co-elute. Of course if they come off in the isothermal portion then that will not be the problem, but still make sure they do not co-elute by running them separately just in case.

Looks like Ethyl Acetate responds about 4-5x better than Chloroform so if they do co-elute then the Ethyl Acetate would overwhelm it. I didn't find a reference to DMF or DMSO on that column so also make sure they do not co-elute with Chloroform.

Chloroform normally isn't one that adheres to Vespel or other constituents of the headspace analyzer or carrier tubing so I would not suspect that, unless it has been contaminated by some high boiling hydrocarbon compounds that might adsorb it.

[Jana P]

Hi James,
Yes I have analysed individual standards. Chloroform is resolved from Ethyl Acetate. I see the same issue with chloroform if I analyse it alone or in solution with the other solvents.
I have made some modifications to the method now but I still have issues with chloroform linearity. I am not sure it I am working at levels that are too low. Although all solutions are reproducible with good RSD between solutions there is variability and no pattern to results obtained.

I have improved things slightly but I don't have the confidence that if I were to prepare the same solutions again it would give me the same results.

I am running out of avenue to explore. Wondering if I would have more luck direct injecting rather than headspace? Although that has its own issues.

[James_Ball]

Any other process in the lab that uses chloroform?

We have to be careful with it since cross contamination can happen very easily and if there is any in the air it could be affecting your standards by getting into it when you prep them. Just something else to look for when dealing with highly volatile analytes.