Trouble with Area of Absorption in HPLC

[fall4p]

Hello to all,

I am running an HPLC using as mobile phase: water (ph=2 with phsphoric acid) / acetonitrile : 70/30. I run the same solution with three injections and I get the area average 4645 mAU, which is very high for this specific solution. I notice the level of water is very low and change it with a new fresh solvent with the same pH. Then I run again three injections of the same and get 4450 mAU, which is normal. Can somebody please understand why? I am thinking baseline drift since the only thing I did was to change the mobile phase. I exclude the chance of the first solvent being contaminated because other solutions I runned with the first solvent showed good response.

Thank you for your help in advance

[tom jupille]

First of all, I am assuming that the sets of three injections gave reproducible areas (i.e., the standard deviation of each set was significantly less than the difference between sets). That probably eliminates variations in injection volume. Since you only mention one peak, I am also assuming that peak is well-resolved from any interferences.

Peak area is a function of three things: the absorbance of the compound, the concentration in the flow cell, and the residence time in the flow cell. You need to overlay a "good" and "bad" chromatogram to get a closer look at what changed. Some possibilities:
- the retention times are slightly different: perhaps the flow was a bit slow in the first set, which would increase the residence time in the flow cell and the width of the peak. You can look at the dead time (t0) to confirm a flow problem. If t0 is correct, then flow was *not* a problem, and any retention change was due to differences in the chemistry of the mobile phase (presumably, the pH); that, in turn, may affect the integration parameters.
- the residence time and peak width are the same, but the "bad" peak is higher: perhaps the absorbance spectrum of your compound changed due to a pH shift in your mobile phase.

Those are the major things that come to mind. I generally recommend that aqueous buffers be prepared fresh daily. Microbes can grow *very* quickly!

[fall4p]

First of all, I am assuming that the sets of three injections gave reproducible areas (i.e., the standard deviation of each set was significantly less than the difference between sets). That probably eliminates variations in injection volume. Since you only mention one peak, I am also assuming that peak is well-resolved from any interferences.

Yes, in fact both sets of three injections gave reproducible areas (rsd% 0.17 and 0.11). The area I was refering to was the Average Area of the three injections on each solution.

- the retention times are slightly different: perhaps the flow was a bit slow in the first set, which would increase the residence time in the flow cell and the width of the peak. You can look at the dead time (t0) to confirm a flow problem. If t0 is correct, then flow was *not* a problem, and any retention change was due to differences in the chemistry of the mobile phase (presumably, the pH); that, in turn, may affect the integration parameters.
- the residence time and peak width are the same, but the "bad" peak is higher: perhaps the absorbance spectrum of your compound changed due to a pH shift in your mobile phase.

Retention times are always the same, but I checked the pH of different compositions of the mobile phase and there some (small) differences. The reason why I thought of checking different compositions of mobile phase is this: If the low level of the Buffer results a different mix of the two solvents (buffer+acetonitrile) then the composition of the mobile phase changes, thus the pH. So I tried different compositions of mobile phase: Buffer/ CH3CN : 70/30 (this is the composition of the method I use), 60/40, 50/50 and 40/60.
the results in pH I find are: 2.58, 2.70, 2.80; 2.89. This is probably the answer..

Those are the major things that come to mind. I generally recommend that aqueous buffers be prepared fresh daily. Microbes can grow *very* quickly!

I use buffers for two days: the day prepared and (if necessary) the next. I support your argument too, but this is a point of discussion and actually I can't find an article or paper that gives strong evidence about that.

Thank you for your comment. I appreciate your help.