Chromatography Forum

Hi,

I'm using a Shimadzu SPD-M40 Diode Array Detector. Scan range 190-500 nm. I focused on 210 and 307 nm for compounds of interest. Mobile phase are 0.1% formic acid in water and acetonitrile with 0.1% formic acid gradient increase % acetonitrile throughout run.

I already read a previous question who used a Shimadzu SPD-M20A. My main point is what is the relationship between UV cut-off and baseline drift. I compared
0.1% formic acid and acetic acid the wavelength from 210 nm to 250 nm. The result was that UV 230 nm of acetic acid (water and MeCN) and UV 240 nm of formic acid have same drifting baseline.

Also, there is no leak and Cell temperature is same as column temperature (40 C). My eclipse C18 column is good condition.

And I checked the UV cut-off of methanol (210 nm) and MeCN (190 nm) in my device. So, I don't understand UV cut-off of formic acid in my device.

UV cut-off is one of those bits of over-simplification that get perpetuated from generation to generation. There is no sudden magic wavelength at which a solvent ceases to be usable, and above which everything is beautiful.

UV/visible absorbance spectra are smoothly changing things. Acetonitrile absorbs at short wavelengths, and even above 200nm it does have some absorbance. The extent to which this matters depends on the size of peak you're looking for. If you have a peak that's going to be 1 AU high, and the baseline drifts a bit you won't even notice, but as soon as you're trying to look at 1 mAU peaks, if you're looking at 210nm, the baseline is going to slope quite significantly on the scale you'll be using to see the peak.

The important things are (1) is the overall absorbance of the solvent mix so high that sensitivity is reduced, and (2) is the slope so bad that your integrator can't find a good baseline? If you can integrate it reliably, it doesn't matter if it's on a slope.

Bumpy baselines are another matter. If you've got a smooth drift, an integrator will cope. If you've got bumps, the integrator will end up finding spurious peaks, and the integrated area of a true peak will depend on where it is, relative to the bumps. Some Shimadzu systems are prone to bumpy baselines. I've dealt with this by subtracting a blank wavelength, but I've heard of others changing the mixer to a larger volume mixer.