Size of negative peaks in RI detection.

[chayyl]

I am using a RI detector to detect the alkanes and mono-aromatics in diesel oil. In addition I have standards which are dissolved in heptane and my mobile phase is also heptane. My standard contains o-xylene, Dibenzothiophene and cyclohexane.

I noticed that there is a negative peak just before the elution of cyclohexane. The size of this peak is about -40000 nRI units. However, when I run a real diesel sample, that same negatve appears but is -7000 nRI units.

I accept that in RI detection, negative peaks are a common phenomenon, however, is there any explanations as to why the negative peaks are of vastly different size for diesel and heptane.

In addition, I am using a manual injector, and I realise that manual switching of the injection valve can give rise to mechanical disturbance that may show up as negative peaks. Given that switching of the valve by the same operator is reproducible, can manual injection produce such great variations if the negative peak is indeed due to a mechanical disturbance?

Many thanks for any reply.

Regards,
Leonard

[HW Mueller]

You know how to find out the scope of your "mechanical disturbance": You just actuate this mechanical disturbance in absence of everything else.
It is not unusual that different samples contain different amounts of unknowns.

[chayyl]

Hi mueller,

Yes. I did a small test to determine the scope of the mechanical disturbance. I manually switch the manual injection valve to the inject position and then start the run without injecting anything. In this way, I assume I will be able to ascertain that the negative peak is indeed caused by a mechanical action of the switching and not chemical in nature.

I did this for several times and monitor the size of the negative peak, and it seems to be pretty consistent for blank injections. When I injected a diesel sample, that same negative peak was vastly different in size to that of a blank. That is what puzzles me.

[HW Mueller]

This is not so surprising: It´s difficult to exactly mimick the injection process without an injection.