HPLC UV Detector Calibrations

[rnelson]

I'm curious how different companies calibrate their uv detectors on hplc. Mainly, how many wavelengths do you check? Do you perform the procedure by making injections, or is your procedure more manual?

I'm considering a change to our in-house procedure for hplc detectors because it is quite different from our calibration for uv spectrophotometers.

I'd really like a wide sampling of what companies are doing, so even if you do exactly what someone else does, please reply so I can get a good representation. Thanks for your input.

[DR]

Many UV/VIS detectors come w/ built-in, sealed Holmium oxide filters that cover the visible spectrum but leave the UV pretty well out of the loop.

We specify 2 options for UV calibration.

1) uses NIST grade potassium dichromate in an accurately prepared set of 4 solutions that are read at 3 or 4 different wavelengths. These readings are used to compare calculated molar absorbtivities to theoretical values.

2) uses a caffeine in water solution to confirm that caffeines local max and min are within one or two nm of their theoretical values (both in the UV range of the spectrum). This is, by far, the easier fo the two approaches. It is not as holistic as the first option, but if all you really need is wavelength accuracy in the UV, it's a good way to go.

[Bruce Hamilton]

As noted above, most manufacturers have internal filters in their systems. They may also have full qualification protocols ( eg using caffeine solutions ), that can provide further assurance on performance.

I would look to the recommended existing qualification protocols, and not introduce further testing, unless those protocols were deficient, when compared to any relevant regulatiory requirements.

I expect detector manufacturers have shown their holmium tests/recalibrations are sufficient over the range of the instrument, and their qualification protocols supplement the internal tests.

When I was in QC, we just followed the HPLC manufactuer's recommendations for qualification and routine performance testing.

My suspicion would be that user selection of detection wavelength amd spectral band width are more likely to produce errors.

For checking absorbance of UV-Vis spectrometers, I actually prefer the convenience of the very-expensive calibrated filter sets ( available from NIST ) as they are so simple to use, with no issues about solution preparation. Not sure where they fit in the regulatory framework though.

Bruce Hamilton