how to deal system peak

[HW Mueller]

DR, what sort of UV/VIS machine do you have?! I have never seen anything that isn´t at least 100X less sensitive than a UV/VIS detector. A lot of stuff looked good in the spectrometer but was lousy in the detector.

[DR]

Any decent UV/Vis should give you some idea about a solvent when you run a scan eg: 300nm-190nm. If not, retry it with a 5cm or 10cm flow cell. You won't get much absorbance, but you will see a (relatively) steep rise start somewhere between 220nm and 190nm. All other things being equal, select the lot whose curve is farthest to the left.

I typically use a Shimadzu 1600

[HW Mueller]

Strangely, there really are still some differences in HPLC solvents on the market which one can see on a spectrophotometer (diffs near 10^-3 AU), especially surprising to me is the variation in ACN. In connection with H2O I think this was presented before: When I want to know whether there are diffs on my detector I put the solvents to be compared, one after the other, into an emptied prep column (100+mL, fitted with an SSI needle valve on the bottom, vakuum or N2 can be plumbed to the top) evakuuate for ~10 min then run them directly through the UV detector with N2 pressure. One can nicely see differences of a few 10^-5 AU.

[anupama]

I saw this type of problem during my Ph.D time. I injected blank HPLC grade different solvent from very good brand & I found a lot of junk.