Chromatography Forum

Any diode array experts out there?

I've got an impurities analysis where we inject the exact same samples on two different brands of LC diode array detectors and get a few components that return higher areas counts on one system.

We tried on multiple systems and are convinced the differences we see in area counts is related to the differences in how these detectors work.

One detector uses a reference wavelength and subtracts that absorbance from the analytical wavelength. The other detector collects a reference spectrum then performs a log calculation to determine sample absorbance.

We've tried quite a few things and I've got a ton of further detail, but didn't want to scare anyone off. The discussion may likely involve the particular detector manufacturers, so I don't mind taking the discussion to email.

Thanks for any input or ideas.

No great surprise there. HPLC is always a relative technique with quantitation based on a calibration plot run on the same instrument on the same day.

If you haven't already done it, try disabling the reference wavelength function (on the Agilent?).

Also check to see that you are running the same bandpass on all systems (this would be equivalent to the slit width on a classical spectrophotometer).

From your description, I'm a little vague as to exactly how you're doing the quantitation (based on a single wavelength? or are you combining multiple wavelengths? or using a "maxplot" to operate at whatever is the absorbance maximum).

Hi,

We have obseved differences between PDAs and VWDs at low wavelength 10nm.
In that case a solution containing 2 substances was injeted in both systems and the area ratios were different (same solution - same wavelength).
Changing to a higher wavelength solved the problem.

Alex

We have obseved differences between PDAs and VWDs at low wavelength 10nm.
In that case a solution containing 2 substances was injeted in both systems and the area ratios were different (same solution - same wavelength).
Changing to a higher wavelength solved the problem.

Which is probably a bandpass issue. At 210 nm (I assume the "10nm" was a typo), you're generally looking at "end absorbance", which has a steep slope. Depending on the exact shape, you can be seeing quite different average absorbance over ± 2 nm versus ±10 nm, for example.