Chromatography Forum

Hi all,

My baselines are fine without any acid modifier, however, when I add 0.1% formic acid I get a steep rise and then a large peak as it hits 95% acetonitrile. I've tried two fresh bottles of 99% and 98/100 from acros and fisher respectively.

I'm hoping that there's nothing in the formic so attribute the shift to RI differences. I'm therefore wondering if anyone knows the different % ratio for acid modifiers in mobile phase A and B to minimise RI baseline disturbances when using fast gradients.

I'd be interested in formic acid, acetic acid and TFA.

If anyone knows of anything else I could do to reduce this effect let me know.

Many thanks in advance,

Formic has UV absorbance out ~230 nm, maybe a little more, so your baseline shift is not a hugh surprise.

When I used trifluoroacetic acid (TFA) with water/acetonitrile gradients, I used 0.1% TFA in mobile phase A (water) and 0.08% TFA in mobile phase B (acetonitrile). This was approximately correct but to achieve the flattest baseline (not perfectly flat) I would measure the absorbance of the initial mobile phase composition and the ending mobile phase composition and then add TFA dropwise to either of these so that the absorbance in the ending mobile phase was about 0.03 less than the absorbance in the initial mobile phase. You can do the same for formic and acetic acid to determine the best amounts of each. Just don't expect to achieve a perfectly flat baseline.

AA

I've seen several reported values for the UV cut off of formic acid, some as you say go up to 230nm. But for instance this site states 210nm. http://www.mac-mod.com/tr/pbprep-tr.html
I’ll probably take a look for myself.

One question:
If formic is present in both mobile phases in equal amounts then there is clearly no change in the quantity of formic acid over the gradient. Therefore, are you saying that formic acid has a much higher absorption in acetontiile than in water? Again, would a change in % of formic acid in each eluent help even this out?

Serpico - I'm not familiar with using gradient for RI detector (as in your title). For RI we most often pre-mix the mobile phase to minimize baseline.

I thought you were using a UV at low wavelength. If you are, in fact, using RI, you will never have a flat baseline with a gradient. Never ever, isocratic only. A differential RI detector (which is most of them) measure the difference between what is in the reference side of the cell and the sample side of the cell, in a gradient the sample side is always changing.

No, this is UV...225nm, so I thought formic would be ok. But the baseline disturbance as it goes to 95% acetonitrile is considerable. But only when 0.1% formic is in the mobile phase. - I've just run it on a DAD and the wavelength would have to be ~250nm to get an acceptable baseline.

Post time increases don't alter the size of the disturbance so I don't believe it to be impurity build up being washed off.

I have heard of different % TFA being used in mobile phase A and B to even out baselines and was wondering if the ratios are widely known for common pH modifiers.

Is this increase a UV absorption effect or an RI effect or something else? I can't work out why it should be either UV or RI. Could someone explain the effect in a bit more detail so I can understand what's going on.

From the information you got so far you could deduce that absorbance is the primary cause. One easily forgets the definition of cut-off point: It is the wavelength for which the UV absorbance approaches 1 (NOT ZERO) for the neat liquid, usually given for 1cm cells (careful!, sometimes a 1mm path is used).

Apologies but I think I'm missing something.

The switch from water to acetonitrile does not produce notable change in absorbance at 225nm. As shown by running the gradient without formic acid.

When 0.1% formic acid is put into both mobile phases it should have a constant concentration throughout the gradient. Therefore, even if there is absorption due to formic acid it should be constant throughout the run - I assume that statement is wrong, but why?

skunked_once already gave you the answer to that. The absorbance of compounds is matrix dependent, especially those of ionizeable (dissociating) compounds like acids.

skunked_once already gave you the answer to that. The absorbance of compounds is matrix dependent, especially those of ionizeable (dissociating) compounds like acids.

And, to complicate matters, just because you add the same amount of formic acid in both the A and the B solvents doesn't mean that you won't see shifts in concentration. formic acid is (slightly) hydrophobic and some of it is probably retained on the stationary phase early in the gradient only to elute at higher organic solvent concentration.

If you want to tweak the separation, prepare a set of 0.1% solutions of formic acid in various ACN/water mixtures (10%, 20%, . . .) and run absorbance spectra on all of them. That will tell you what's going on with wavelength. If possible, try to find an isosbestic point in the spectra and run at that wavelength. Then do a series of experiments with less than 0.1% formic in the B solvent until you get the best baseline (you'll never get it perfectly flat).