anyone know is there a general cause for negative peaks in RP HPLC, say, water/ACN as MP, UV detection at 214nm, for pharm. samples. the MP have already low UV cutoffs, and the detection wavelength is quite low, i suppose anything absorbes should give a positive peak, unless there is another solvent residue in the sample which has even lower absorbances than water and ACN? like what?

For the most part (although with some detectors refractive index effects could create negative peaks), negative peaks are due to impurities in the mobile phase being had a higher concentration than what is found in the sample. Regardless of the quality of the solvent there are likely to be low-level impurities present and work at 214 is certainly more than sufficient to visualize these. The UV cutoff specification tends to give one an undue sense of quality from an analytical point of view. For example, for a high-quality UV grade acetonitrile the cutoff is typically 190 nm. That might sound good but the cutoff is defined as an absorbance of 1! Looking at the handbook for HPLC solvents from Burdick & Jackson, they list the absorbance at 205 as 0.1 and the absorbance at 225 as 0.01. Presumably, the absorbance at 214 is somewhere in the vicinity of 0.05. If one assumes that this absorbance is due to 1-3 major impurities which are most likely not in your sample, then you should certainly expect to be able to see negative peaks where each of these solvent impurities elute if you are working at a sensitive range on your detector.

What is the detector manufacturer ? I have seen the Agilent diode array detector give a negative peak wherein the method had a reference wavelength which was at/near the UV cutoff of the component eluting (ie toluene). In this caase, moving the reference lambda out to 500 nm corrected this...

thanks, the detector is waters PDA, but i think Chris's comments make sense.
thanks again

Noname,
For light detector a plug of solvent moving down the column can make a lense in the detector because of its shape and the difference in refractive index of the plug and mobile phase. This lense can bend the light away from the detector and cause a negative peak.
Jim

I thought that a negative peak is an increase in light?!

Modern detector flow cells are designed to compensate for and remove refractive index effects. The Waters PDA (the detector in question here) is no exception. So although Mr. Gorum's comments may be theoretically correct, they are, for all practical purposes not a concern in this case.

What is happening here is that the mobile phase
has a component which (1) absorbs light at the
wavelength selected and (2) has a given retention
time which is greater than ACN i.e. ACN cannot
displace the component from the stationary phase.
The injection therefore dilutes the mobile phase
wrt to the component: creates a mobile phase
segment that has a lower absorbtivity than the
rest of the mobile phase. The diluted mobile phase
segment moves thru the column with the same
retention time as the component but the absorbance
is less than that for which the baseline was
balanced therefore producing a negative peak

Seems that there are some puzzling answers here, except maybe Chris Pohl´s. But, your negative peak could have as simple a cause as having more water in the sample than in the mobile phase. Water absorbs less than H2O/ACN.

Why does impurity make peak?

Most of the discussion here is about the initial negative peak observed in a HPLC run. What about negative peaks in the middle of a run? I have noticed them to be prevalent when you have a mis-match between the mobile phase and the solvent used to prepare the sample (especially when the solvent is acidic). Are these simply impurities eluting late which has lower absorbance than the mobile phase?

Last anonymous: Anything that dilutes the MP and absorbs or scatters less than the MP(or reduces scattering of MP) has to give a negative peak if it moves through the detector, anytime.