I am using HP1050 with diodarray detector and monitor at different UV wavelengths simultaneously. When I overlaid signals after the run I noticed a slight peak shoulder appeared at only one of the selected wavelengths. Does this necessarily mean that there is a hidden impurity under the peak or is it possible because of the non-linearity of the detector signal at one of the wavelengths? Diodarray peak purity criteria did not help here.

Could you tell, at which wavelength the shoulder appeared? Was it at 200-210 nm? It looks rather as an impurity, as a detector nonlinearity to me.
There are ways to find out if it is an impurity:
try to change gradient time by a factor 3, the pH of 0.5-units up and down or use a 20°C higher temperature. Your peak might be moving away, so you can quantitate it.

Imre Molnar

Unless you experiencing detector overload at that wavelength there is probably something under the peak. You can test detector overload by diluting the sample and noting the peak shape. Diluting not only effects detector overload but to a small degree increases column efficiency thus the impurity may be further resolved.