[dolphinsea14]

How to choose optimal wavelength in spectrophotometry if wavelength-absorbance graph is complex, has more than one absorbtion peak? Should I choose the peak at maximum wavelength, the one that there is greatest difference in absorbances for 2 wavelengths, or the broadest peak?

[wildfyr]

What are you trying to use it for? Optimal for what? If its just for concentration then any peak would work as long as there isn't an impurity or solvent peak that overlaps. The strongest peak is preferred in the absence of anything else, because you can see it an the smallest concentrations.

[dolphinsea14]

It was a sample question on test, we were given a graph with few absorbtion maximas, the one for the greatest absorbance was very narrow, another for a bit lower absorbance was broader, and there was one with greatest difference between 2 absorbances. We were asked which wavelength would we choose and why?

[Babcock_Hall]

It is not a simple question, because it depends on several factors.  Suppose you were asked to focus on adherence to the Beer-Lambert law.  Which peak would potentially be best, and what other factor comes into play?

[dolphinsea14]

Let's consider a general case.
If a monochromatic source of radiation is used, can I choose the highest but narrowest peak, or should I choose a bit lower but broader one? I read Beer's law applies for radiation with single wavelength, for polychromatic radiation, the peak should be broader, so that molar absorbtion coefficient is constant in the interval of wavelengths that is used, and the deviation from Beer's law is low.
But what about monochromatic radiation? And in general case?

[Babcock_Hall]

I think that you understand the issue that I was hinting at, namely that when radiation is not monochromatic (for example, when the band width is broad), a narrow peak will show a deviation from Beer's law.  If it is essentially monochromatic, then I don't see a problem with using a tall narrow peak.

[Corribus]

First, there is no such thing as monochromatic light. Even lasers have a finite bandwidth.

All things equal, the best situation is if the bandwidth of your absorption band significantly exceeds the bandwidth of your incident light. Or, put another way, that the absorption feature is locally "flat" over the region spanned by the incident light - a situation that is easier to accomplish if the spectral bandwidth of the incident beam is small (less polychromatic). If the bandwidth of your absorption band does not significantly exceed the bandwidth of your incident light, you will observe deviations from the Beer-Lambert law because you will have complex and nonlinear interactions between the incident beam, which consists of different wavelengths of light with different powers, and the absorption band, which consists of different wavelengths of varying molar absorptivity. The same thing is observed when, for example, you are trying to measure the absorptivity at the side of an absorption band, where the extinction coefficient of the absorption spectrum changes rapidly as a function of wavelength across the spectral area of the incident beam.

There’s a really nice Applet that allows you to play around with this effect and see how varying degrees of spectral bandwidth affect deviations from ideality.\

http://195.134.76.37/applets/AppletBeerLaw/Appl_Beer2.html

(There's also an instrument issue for broad spectral bandwidths - and that is that point measurements using a simple photodiode are often not calibrated to take into account different sensitivities of different light wavelengths.)

The originally posted question isn’t a very good one, because there are also other considerations than just the deviation from Beer’s Law that you may have to take into consideration, most notably the concentration range over which you are likely to be measuring. Although it would seem to make sense to always pick the strongest peak, in some situations it is actually better to pick a weak peak. For example, if the concentrations you will be measuring are very large, picking a strong peak (strong meaning high molar extinction) could bring you into a range where the instrument doesn’t perform well – e.g., where the absorptivity is > 1 OD. In such a situation, a weaker peak would be a better choice. (You could also simply dilute the sample or use a smaller path length, but there are drawbacks to both of those options as well.)

[Raayou]

When analysing  wavelength data,  inspect them to see whether  that the absorbance  peaks have the proper looking  shape and size.A more correct wavelength would produce  series of absorbance signals. You might wanna look at the absorbance  from the blank because it has to produce a signal with a relatively low intensity and the absorbance signals for the calibration standards and the  samples should produce a symmetric and Gaussian sort of like shape with a single peak and each peak should exhibit a proportional change in magnitude that will match the difference in concentrations. 😀.

[dolphinsea14]

First, there is no such thing as monochromatic light. Even lasers have a finite bandwidth.

All things equal, the best situation is if the bandwidth of your absorption band significantly exceeds the bandwidth of your incident light. Or, put another way, that the absorption feature is locally "flat" over the region spanned by the incident light - a situation that is easier to accomplish if the spectral bandwidth of the incident beam is small (less polychromatic). If the bandwidth of your absorption band does not significantly exceed the bandwidth of your incident light, you will observe deviations from the Beer-Lambert law because you will have complex and nonlinear interactions between the incident beam, which consists of different wavelengths of light with different powers, and the absorption band, which consists of different wavelengths of varying molar absorptivity. The same thing is observed when, for example, you are trying to measure the absorptivity at the side of an absorption band, where the extinction coefficient of the absorption spectrum changes rapidly as a function of wavelength across the spectral area of the incident beam.

There’s a really nice Applet that allows you to play around with this effect and see how varying degrees of spectral bandwidth affect deviations from ideality.\

http://195.134.76.37/applets/AppletBeerLaw/Appl_Beer2.html

(There's also an instrument issue for broad spectral bandwidths - and that is that point measurements using a simple photodiode are often not calibrated to take into account different sensitivities of different light wavelengths.)

The originally posted question isn’t a very good one, because there are also other considerations than just the deviation from Beer’s Law that you may have to take into consideration, most notably the concentration range over which you are likely to be measuring. Although it would seem to make sense to always pick the strongest peak, in some situations it is actually better to pick a weak peak. For example, if the concentrations you will be measuring are very large, picking a strong peak (strong meaning high molar extinction) could bring you into a range where the instrument doesn’t perform well – e.g., where the absorptivity is > 1 OD. In such a situation, a weaker peak would be a better choice. (You could also simply dilute the sample or use a smaller path length, but there are drawbacks to both of those options as well.)

Thank you very much, Corribus, you explained everything very well. Also I found the link very useful. I really appreciate your help.
All the best