Hadeia,

Changes in baseline absorbance are normal in gradient methods especially when gradients are changed rapidly (e.g. 5% ACN to 55% ACN in 2 min) and observed wavelengths are in the shorter UV range. You can test to make sure this is the cause of your peak (which is probably just the baseline) by running the gradient without any injection and observing the chromatogram. The peak should still be there and be of the same intensity without an injection. This should not impact quantification as long as the peaks of interest are well resolved from the baseline event. The peak can be minimized by slowing down the rate of change in the gradient or by quantifying at a longer wavelength.

Benjamin Chrisfield

As you said the peak is related to the baseline and it becomes broader and minimized by slowing down the rate, but I can not use slower rates as the run time reached about 30 minutes, my problem is in method reproducibility it works good for several injections after that either a cut off peak or a negative peaks appears and I think the problem is related to this baseline peak

Ghost peaks cam come from many places- solvent, the solvent used to dissolve the sample, or other places, sometimes even a previous run. You can test if the mobile phase is the source by increasing the equilibration time before injection, and seeing if the area under the peak increases.