Has anyone ever encountered negative deflections in the baseline outside the solvent front, on an isocratic system?? We are currently using 55% CH3CN 45% 0.01M KH2PO4 pH 3.0 on an Inertsil ODS-2 5um column at 250 nm. Our peak of interest elutes at 8 minutes. The negative deflection appears at 8 minutes. This is not a constant phenomena. First, we were encountereing an interference from the blank with the peak of interest. I determined the interference was coming from the water we are using. So, unable to obtain better quality water(B&J bottled H2O) we were going to subtract the interference out. So before each sample we would inject a diluent blank(55% CH3CN/45% H2O). When we did this, instead of the intereference we were getting the negative deflection instead!! I checked the diluent, put on a new column and even changed HPLC systems!! Still there. The only thing that has remedied the situation is new mobile phase. Anyone know why????

Dear Deb,
I have an inkling. It sounds to me like you are doing some vacancy chromatography. In vacancy chromatography, the "analyte" might be added to the mobile phase and then an injection of pure solvent made. The "analyte" peak then appears as a negative peak, assuming it has a higher absorbancy compared to the pure solvent. The negative peak that you observed when you injected the diluent blank might have been a component that had inadvertantly been added to the pumped mobile phase. If you can inject a mixture of diluent plus exactly the same amount of that component you will see a perfectly flat baseline. Vacancy chromatography is actually a nice QC method for common analyses since retention time imprecision is not really important. One reference is Guiochon et al Anal Chem 1990, 62, 923. Good luck!

Marc

Another possibility.

What you probably see in your chromatograms is a system peak due to mobile phase absorption at chosen wavelength. System peaks are usually due to poor quality of mobile phase constituents like water or acetonitrile as well as reagent grade or contaminated additive ( KH2PO4) or pH adjusting solution.
If you are unable to obtain new chemicals and still want use the blank subtraction alternative, I would recommend to inject two blanks in the row and use the second blank for your blank subtraction calculation.
Why? Because, while you first equilibrated the system, some junk from the mobile phase was loaded on the column. In your second injection, most likely, less junk was loaded onto the column since the interval between injections was much shorter then the time it took to equilibrate the system. Now, when you subtract your blank, you get a negative peak because, the system peak in the blank is larger than the one in the sample. During the second injection of a blank, things should be more normalized in relation to the following injections, that you should see a very minor or no disturbance to the baseline in the area of a system peak.

It appears to me that your case is due to impurity present either in the water, phosphate(less likely in most cases), the 85% phosphoric acid used for pH adjustment (are you using HPLC grade? It's necessary.) or contaminated mobile phase filtration system. Your diluent blank is not or less contaminated and with a less absorption than the mobile phase, therefore you experienced a negative peak. To this end, you can not subtract this blank. Nevertheless, this is not a good practice and is subject to query by the regulatory agency. Try to inject an aliquot of the mobile phase to see if you are getting a negative peak or not. If you don't have any peak, you know the answer. If you still see a peak, further investigation is needed.