Baseline Drift UPLC

[tjk123]

I am noticing a baseline drift negative when using a gradient elution on UPLC

MPA: 80:20 (1.0% TEA Aqueous:ACN)
MPB: 20:80 ((1.0% TEA Aqueous:ACN)
btw, all high quality reagents

Temp. 30C
Detection: 280 nm

Baseline starts out flat during isocratic then proceeds to ramp lower and then remain flat when isocratic again.

Neither ACN nor TEA absorb much at 280 nm so this maybe the "norm" of the UPLC.

I am just perplexed that when I run this same method on an HPLC (with gradient volumes identical), the baseline treads flat during isocratic then proceeds to ramp higher then remain flat again when isocratic again.

I have monitored the pressure on the UPLC and ironically, the baseline has the same shape/pattern. Any correlation?

[tom jupille]

Most likely different responses to the RI shift.

[tjk123]

Is there any way to "fix" this issue?

Also, how exactly is differing RI affect a UV detector, especially with solvents that are practically with UV footprint in the region.

I saw another thread that suggested it may be a result of a dirty flow cell or improper mixing but I'm not sure it affects me since the thread related to TFA.

[tom jupille]

If you have a diode-array detector that lets you set a reference wavelength, that can help, but a certain amount of RI sensitivity is unavoidable. Because you cannot perfectly collimate a light beam (absent a laser, of course) there will always be *some* light entering the flow cell at an angle, which means that the RI of the cell contents will have *some* effect on the light throughput. If the flow cell is out of alignment, the problem is exacerbated. Whether the effect is positive or negative can vary from one instrument to another.

If it's not too bad, just ignore it, in many cases detectability is hurt more by garbage peaks than by drift. If it *is* causing a problem, try enabling a reference wavelength as suggested in the previous paragraph. Other than that, making sure the flow cell is properly aligned can minimize the magnitude of the drift.