SPME, unwanted peak shapes with splitless mode

[ChrisT]

Dear GC-users,

we're analyzing food samples via SPME(Headspace)-GC/MS. Typically SPME is performed splitless, but the splitless mode makes that the peaks of the first 8 minutes have strange peak shapes (chair-like, broad, large tailing). If we switch to a split mode (1:10) during injection, the signals improve massively, but we lose intensity. Why can't we measure splitless as the majority of SPME-users? What is wrong with our settings? I would highly appreciate your opinions.

Kind regards, Christian


Our current "splitless" conditions for the injection are:
Splitless for 45 s, then 1:25
Desorption time: 180 s
Inlet T: 270 °C
Liner: 1 mm diameter (smaller not available for our machine)
Fiber: DVB/CAR/PDMS (2 cm stableflex)
Column: DB5-MS, 30 m, 0.25 mm ID
Gradient: 35 °C to 240 °C with 6 °C/min.
Flow: 1 mL/min He

We tried to solve the issue by
- increasing the inlet temperature (270 °C instead of 250 °C): no clear effect
- increasing and decreasing the time of the splitless mode in the beginning: no clear effect
- changing the column to a DB-Wax: no clear effect on peak shapes.

[James_Ball]

Is the initial hold of the oven temperature longer than the 180s desorption time?

If the oven ramp begins at time zero, then you are ramping the oven as the analytes are being loaded onto the column, which is going to cause tailing and broad peak at the beginning. You need to allow the analytes to focus on the head of the column during the loading time, then hold a little extra before ramping the oven temperature.

The split injection helps because the flow around the fiber is higher and the analytes come off more quickly and require less time to focus on the head of the column.

Try splitless with a longer oven hold, maybe even 2.5 minutes, and a cooler initial temperature to improve the focusing.

[ChrisT]

Your hints helped!
Starting the split after the desorption finished (after 3.5 min), massively improved the signals. The result is better than with a constant split of 1:5. As expected the signals are more intense for the splitless start. We will fine-tune the method now.

Having no gradient for the first minutes (also 3.5 min) also improved the signals.
We didn't lower the start temperature as the cooling might get difficult in the summer time. So, I can't give feedback on this.

[James_Ball]

I'm glad it helped.