Hi Thomas,

as you know, the detector does not measure absorbance, but rather the (in this case transflected) light (or signal) from the sample and calculates the ratio to the reference light measurement. If the absorbance at a certain frequency is too high, depending on the original light levels, the Signal to Noise level may become so low that the Noise dominates, and the Absorbance values appear to saturate.

In fact, in some cases, the reflection from the top surface of the sample can dominate, resulting in a "dispersion artefact" normally associated with scattering effects.

Article Reflection contributions to the dispersion artefact in FTIR ...

Hi Hugh,

many thanks for your contribution! I agree, one can easily think that saturation could become a problem, and I consider your last comment as key towards the final answer.

Hi Thomas Mayerhöfer - the detector does not actually saturate, in the conventional sense, but rather the signal is reduced to the level of the noise, and then, this affects the calculation of the Transmission....... Noise/Reference, and therefore the Absorbance.

Sorry, Hugh, yes, I agree, "saturation" is not the correct term. Let me rephrase my last response:

I agree, one can easily think that excessive noise due to diminishing light intensity could become a problem, and I consider your last comment as key towards the final answer if this would be the case or not.

Thomas Mayerhöfer the same is very commonly observed in UV/vis spectroscopy, particularly in the UV tail, where the light source has diminished intensity.

In either transmission or transflection (FTIR or UV/Vis), the situation can be further complicated in samples which are spatially inhomogeneous over the sampling area, such that portions can "saturate", whereas others act as "pinholes", resulting in a spectrum that does not appear optically too think, but can exhibit "saturation" effects.

Hugh J Byrne So you think that for a 20 micron thick layer in the absence of interference effects one should observe in transflection spectra regions where the intensity that reaches the detector is so low that the detector signal just consists of noise? Let's make things simpler and assume that the same layer with a thickness of 1 micron has a band with a peak Absorbance = 1.

Sorry Thomas Mayerhöfer - I was just talking in general terms.

Hugh J Byrne Have you ever come across a spectrum of a thick layer on a metal or low-E glass (be the layer homogeneous or not) that displayed the effect?

Which effect, Thomas Mayerhöfer ?

For sure I have seen examples in which the surface reflection spectrum dominates over the transflected spectrum (we showed this in the first Bassan paper), also examples in which "saturated" peaks appear, but the average absorbance is not very large.

Hugh J Byrne The effect that absorbance > 3 for a transflection spectrum of a sample of a thick layer on a metal in the absence of interference effects. By the way, the domination of a "surface reflection spectrum" over a transflection spectrum is absolutely natural for thicker layers, since the reflection from the surface is what remains in the limit of infinite layer thickness (and this is what actually limits absorbance). If you use wave optics, you anyway always take this properly into account.

Thomas Mayerhöfer Yes, absolutely, but a lot of people were not aware of its contribution to the so-called "dispersion artefact, and it is neglected in most of the treatments of scattering in transmission -the so-called Resonant Mie effect.

In truth, I don't have a huge amount of hands on experience of FTIR measurement, and less so in transflectance, except looking at the results on others. In any absorption spectroscopy measurement, a general rule of thumb is that you should begin to be nervous once the absorbance approaches 2 (T=0.01). The reliability of measurements in and beyond this range becomes very much dependent on the quality of the spectrometer (the amount of stray light, dark noise etc., the strength of the light source in the region of the absorption). "High end" UV-vis spectrometers quote an absorbance range up to 6, but they use double monochrometers and internal filters to minimise stray light, and even then, this is only achievable (if at all), in the centre of the spectral range, where the light source, grating efficiency, detector efficiency etc. are maximum.

Ok, so I think it is time to reveal that in transflection in absence of interference effects absorbance can never reach a value of 2, in most cases not even 1.5. Why? Because in the limiting case of a very thick layer light is no longer able to reach the metal, be reflected, leave the surface and contribute to the reflectance. Therefore only the reflectance from the surface remains. This reflectance may be comparably weak (index of refraction = 1.5 gives 4 % reflection for normal incidence, which leads to A = -Log10(R/R0) = 1.4 - certainly using absorbance in this case makes no longer any sense), but in any case enough light will reach the detector so that any problems as described above will be avoided.

For more details: Article The electric field standing wave effect in infrared transfle...

Yes - but the reflected spectrum depends on the real component of the refractive index, and so has a very different form, so people should beware.

Transmittance spectra also depend on the real part of the complex refractive index (but admittedly less obvious). As I said, using absorbance for reflectance spectra makes no longer sense. But with a little trick (which I won't reveal for the moment) reflectance spectra can be quantitatively evaluated just as simple as transmittance/absorbance spectra.

Here is the little trick I mentioned in the previous reply: Article Infrared Refraction Spectroscopy