QL very different on different systems

[Mattias]

I am right now validating an LC method on a Waters Alliance with a 2487 dual-wavelength detector. As part of the SST, a 0.10% solution is injected and the S/N must be >10. On the dual-wavelength detector the S/N was over 100.

During intermediate precision, a Waters Alliance with 2996 PDA detector was used. The analyst only got a S/N of 8 (SST failed). I have now been running around the lab with the same bottle of mobile phase, column and sample and it seems to be true difference:

Waters Alliance 2487 system 1: S/N : 112
Waters Alliance 2487 system 2: S/N : 90
Waters Alliance 2996 system 1: S/N : 8
Waters Alliance 2996 system 2: S/N : 21
Agilent 1100 (DAD): S/N : 5

Can it be true that a dual wavelength detector has a 5-10 times better QL?

[Bluesman75]

I don't know the Waters HPLC systems, but have you checked lamp age and intensity.
Are the peak responses comparable or is it just the baseline noise that is different?
Also are the detector cell path lengths the same?

[Mattias]

I don't know the Waters HPLC systems, but have you checked lamp age and intensity.
Are the peak responses comparable or is it just the baseline noise that is different?
Also are the detector cell path lengths the same?

All systems have analytical flowcells (10 mm). Have not checked lamp age (not possible on Alliance I think). It is the baseline noise that differs, not the peak height.

[Bluesman75]

The first things I would check are the lamp and if the flow cell windows are dirty in that case as you have already ruled out mobile phase.

[AA]

I am not overly surprised by this result.

The 2996 is a pretty old design (it was replaced by the 2998 some years ago). The 2489 is a much newer design.

The noise spec of the 2996 are 1.5X10e-5
The noise spec of the 2489 is 5X10e-6

As well, being old, your 2996 optics are very likely fogged up after years of high intensity UV light passing through. This means less total light throughput which leads to overall higher noise.

AA

[Mattias]

The first things I would check are the lamp and if the flow cell windows are dirty in that case as you have already ruled out mobile phase.

Have changed flow cell on the "S/N = 8" instrument, but it did not make any difference. Delta psi is very low, so the pump is working fine. I guess next step is to change the lamp and after that to call for service of the optics.

I am not overly surprised by this result.

The 2996 is a pretty old design (it was replaced by the 2998 some years ago). The 2489 is a much newer design.

The noise spec of the 2996 are 1.5X10e-5
The noise spec of the 2489 is 5X10e-6

As well, being old, your 2996 optics are very likely fogged up after years of high intensity UV light passing through. This means less total light throughput which leads to overall higher noise.

AA

So even if the instrument is working as it should, I can expect three times higher noise of the PDA (and hence three times lower S/N)? I have to check what QC is using...

[unmgvar]

Mattias you cannot compare the s/n of a PDA to that of standard UV detector
the optics are different the noise in the system is very different

there are several key characteristics affecting the noise levels and in general as a rule of thumb a PDA will be 6-8 times noisier than a UV. you can try to improve results, but you might never get them as good as for with UV detector


also between instrument of the same type you could have effects of column, tubing connections, tubing dwell volume, sampler injection offsets and flow cell that can give you different results

overall you are in pretty good shape with your DAD's
and the difference between your systems is pretty good

[LC_labrat]

Detector energy does affect S/N, you can check the energy level in the diagnostics screen. At 50 it will not calibrate, it is about 100 for new lamps. Causes of noise vs light intensity are lamp, dirty flow cell and optics. Optical resolution for a 2487 is run at 4.8 nm, it should match PDA to get similar results.

[prepcolumn]

For Alliance systems there is around ten times S/N difference between dual UV and PDA detectors.

[mac]

Those results wouldn't seem strange to me OP. The dual detector will be much more sensitive than the PDA. I've found the same even on the more modern HClass systems between TUV and PDA even though the difference is only 3 or 4 times with those.