Chromatography Forum

I normally set the reference wavelength on my DAD to 360. I thought that any higher wavelength would use the tungsten lamp as reference and thereby increase noise. Does anyone know about this?

Thanks,
Marc

Which lamp is used for the reference wavelength is not important. The important thing is that the reference wavelength doesn’t overlap the detection wavelength, but still detects the baseline fluctuations caused by the absorbance, light scattering and refractive index effects of the mobile phase.
I think more info is available, as we have discussed the topic earlier – a couple of years ago.

Best Regards

For sure the wavelength needs to be one at which the molecule does not absorb. Which lamp is used is important because of the random fluctuations you mentioned. If both sample and reference come from the same lamp then the fluctuations largely cancel. If two different lamps then they don't cancel and the signal is more noisy.

Anybody else know?

I’m afraid you’ve got it wrong. The source/lamp is not important – the cancelling is performed software wise and thus independently of the lamp. Another prerequisite is the fluctuation should be detectable both at the working and the reference wavelengths, which is mostly true in the case of light scattering and refraction.

As mentioned previously you might like to make a search f. ex. here on the forum and you’ll find more info on the matter.

Best Regards

Dancho, how does the software know how the UV lamp fluctuates if the apparatus monitors only the Vis lamp? The Linear detectors which I have split the light at the set wavelength, if I understood the manual, so that the wavelength used for detection is corrected.

Hi Hans,

Glad you asked

OK, that’s how it works:
The working wavelength (beam) is still split just like on any other detector/spectrophotometer. The role of this particular design is meant to correct for lamp energy fluctuation and to give the possibility of zeroing the signal. So, the measuring beam goes through the flow-cell whilst the reference (the split) beam is redirected round the flow cell and to the PMT. These two beams are of the same wavelength and in fact originate from the same source/lamp. The reason that they don’t cancel each other and thus cancelling the analyte’s absorption is precisely the fact that the reference beam does not traverse the flow-cell. That’s the traditional referencing model. In the case of Agilent’s 1100 and 1200 detector models there is a second reference beam at a different wavelength to the analytical ditto. This reference beam traverses the flow-cell as well, but because it has a different wavelength than the analytical one (and does not overlap it) it doesn’t cancel the analyte’s absorbance and thus the peak is not affected.
Finally because effect like light scattering and refraction are less dependant on the wavelength compared to the molecular absorption there is a reasonable chance for the fluctuations caused by gradient and other artifacts to be detected both at the working wavelength and the one chosen for the reference wavelength (regardless of their origin – lamp wise).
So, the software compares these two signals and sets the resultant/derived signal to zero, thus cancelling the detected fluctuations (e.g. baseline rise or whatever due gradient etc.)

Hope the above is not too sleep-inducing

Best Regards

Well, it reiterates what has been said before: It is best to shut it off.

Actually I didn't mean to offer an oppinion on whether or not the facility should be used. I just tried to explain the mechanism.
In some cases - if utilized wisely - it can add a value.

Best Regards

The original poster mentioned the Agilent DAD, which I assume is a single beam instrument.

My understanding of the DAD optical system is :
Vis->(1)->UV->(2)->F?->[Cell]->(3)->{S}->\G\->DAD

1,2,3 are lenses,
Vis = Tungsten Lamp
UV = Pass through Deuterium Lamp ( Vis energy combines with UV )
F = Optional Holmium Filter switched in to calibrate wavelength
[ & ] are windows to the cell compartment
{S} = Variable Slit
\G\ = Monochromator Grating
DAD = 190-950nm Diode Array Detector

If the above is the Agilent DAD arrangement, it would suggest that the cell sees all the energy, and any selected reference would be subtracted electronically, along with other manipulations such as dark current signals etc.

The DAD sees the whole spectrum at each sampling interval, so the "reference" would be a selected subset of the spectrum. I'm not sure whether the use of the reference affects the slit setting, it may, but it shouldn't compromise the desired data.

I could be wrong, but single beam instruments would put more energy into the sample cell, so I'd expect a DAD to use such a system.

Bruce Hamilton

Here is an opinion from my boss who teaches courses in how to use and optimize DAD detectors (many brands including Agilent), we strongly suggest that 99% of users turn the Reference Wavelength feature 'OFF' at all wavelengths UNLESS you have a specific application where it may be useful. The 'Reference Wavelength' feature, as supplied by Agilent and many other companies, has several negatives: (1) They supply very poor default values for the Reference Wavelength and Bandwidth settings in the software. I can not tell you how many scientists we see whom use these 'default' settings in ALL OF THEIR RUNS ! Most people never both to read up or understand how this feature works and that can be dangerous. There is an assumption out there that the manufacturer put these values in for a reason. I am sure that they never thought people would just use them w/o changing them to fit their sample first. We have asked them to provide a default setting of "OFF" for this feature for over twenty years, but they have not complied. (2) By definition, the Reference feature subtracts RAW DATA through the software and only outputs one final result signal. You will never know what the original signal looked like before the reference wavelength was subtracted from your raw data. This is unscientific and really bad from a validation point of view. (3) Most people select Reference wavelengths and Bandwidth values that make no sense (again, because they do not understand how to use the feature) resulting in data that is of poor quality (or misleading).

If you want to use the "Reference Wavelength" feature on a DAD:
We recommend setting up two different signals in the settings box; Signal #1 & Signal #2. If you want to experiment with a Reference signal to improve the baseline of your chromatogram, then analyze the sample using a proper wavelength, called Signal #1, then select Signal #2 at a different wavelength and bandwidth to use as a reference wavelength (at a point where a prior DAD SCAN found no other peaks absorbed). When the analysis is finished, you can easily use the ChemStation software (or any CDS) to manually subtract the reference signal from your analysis signal to obtain the result "with Reference applied". This method preserves the raw data of both signals so you can determine if there is any improvement to the final result. You also have a record of the actual signal as well. This method preserves the data and insures you do not miss something that may be subtracted out using the standard Reference Wavelength feature.

Just to update this very old post...
Here is an accurate description of both how the DAD Reference Wavelength feature works (it has noting at all to do with the lamps) and why you should not use it in method development or in any lab that operates under any type of regulated environments.

"REFERENCE WAVELENGTHS (as used in HPLC UV/VIS):" https://hplctips.blogspot.com/2011/03/r ... -hplc.html