negative peak

[jduan]

Dear experts out there, I have one question about negative peak.

I ran HPLC gradient for a small molecule on Agilent 1100. I repeated the experiment four times, three times worked fine, as I expected. However, one time, I saw a huge negative peak.

Afterward, for whatever the experiments I performed, the system worked fine. It looks to me the negative peak randemly appreared. I am not sure what happened. Will that be possible that the bubble might go into the DAD detector and caused the problem, or something was wrong with the injection?

I just wonder whether there will be experts out there who can tell me what may be the reason for the appearance of the negative peak.

Thanks so much!!!

jin

[tom jupille]

Is the size of the negative peak wavelength dependent? If not, then an air bubble or a lamp malfunction are possibilities.
If the peak is wavelength dependent, then it may be a "system peak" caused by something in the mobile phase which is not present in your sample injection. If a system peak is also a "late eluter", it can emerge only after several gradients have been run; and if you are not doing a lot of automated injections (with constant time between injections), the resulting negative peaks can seem to appear randomly. If that's the case, dissolving your sample in the mobile phase should help.

[HW Mueller]

Is there really evidence for a negative peak due to an air bubble? I have done a lot of gas chromatography with my HPLC equipment, but have only seen positive peaks in relation to any form of gases in the detector cell. One can imagine that a gas bubble, if it causes little or no scattering (how could that be possible? Strange coincidence of no refraction at the interface?), could cause a negative peak since there might be less absorbing material in the cell.

[jduan]

Is the size of the negative peak wavelength dependent? If not, then an air bubble or a lamp malfunction are possibilities.
If the peak is wavelength dependent, then it may be a "system peak" caused by something in the mobile phase which is not present in your sample injection. If a system peak is also a "late eluter", it can emerge only after several gradients have been run; and if you are not doing a lot of automated injections (with constant time between injections), the resulting negative peaks can seem to appear randomly. If that's the case, dissolving your sample in the mobile phase should help.

Thanks so much for the response. First, the size of the negative peak is not wavelength dependent. And also, I ran a lot of sample continuously by automatical injection. There is no negative peak appear periodically. It seems just happened suddently to that run. I think air bubble or a malfunction of the lamp is more possibly the reason.

I have never really met a real case of negative peak exactly caused by the bubble in DAD detector. I only heard people talk about that.

thanks again for all the answers,

jin