We work with biological samples (e.g., bacterial spores; dried, as powder) and we get absorbances of 0.1 to 0.2 with are perfectly suitable for further analyses ( for methodological details see Brandes Ammann & Brandl, BMC Biophysics 4:14, 2011). Generally, the higher the pressure you apply on your sample by the ATR unit, the higher the absorption.

There is no minimum to get a nice clean spectrum. It all depends on the S/N ratio, contact of sample with crystal surface, number of reflections of your ATR crystal and sample thickness.

I got results between 0.01 and 0.3 depending on the sample of interest and number of reflections. There are single reflection ATR units, ATR's with 3 reflections up to the classical one with 25 reflections.

For a powder, 0.35 absorbance is fine.

Regarding your number of replicates:

If you are developing a methodology for characterization make sure you analyze every sample.

If you are following IR intensity for certain correlation, statistics won't help...in ATR-FTIR position of bands is most important to be reproducible not intensity.

You can bet better intensity with thicker samples, higher pressure, better optics, better optical alignment...there are too many factors involved.

So you can see what bands are present and in certain application, what bands are formed or disappear depending on certain interactions.

Dear Emrulla, In some experiments I did with FTIR in gas phase, I had less than 0.005 absorbance units and we made more than 10000 scans with FTIR instrument at a resolution of 1 cm-1, to prove that it is reproducible and to get a better Signal/Noise ratio... To increase the signal (absorbance), we used a multipass system coupled to the spectrometer. But I don't know your experiments so I have no idea if it is compatible...

For example, you can see Fig. 5 in Cameron et al. Appl. Spectrosc. vol. 36 (1982) 245-250. This plot shows dependence of uncertainty in IR spectroscopic parameters for a given band. It stands for both ATR and transmission measurements.

Thus, it is evident from this plot that the errors are minimal if the absorption of the observed band is between 0.2 to ~1 a.u. Accordingly, your band fall in this optimal range and you can measure it accurately.

For quantitative spectroscopy, the rule-of-thumb, has usually been to achieve a signal-to-noise ratio of greater than three.

Dear Emrulla,

You need to have a clear spectrum, it's not necessary a spectrum with higher peaks (mid or far region)! Your value (0.35) is a good value. About the number of samples, depends on what investigation you want perform! If comparisons, guarantee a minimal number of equal samples to have statistics! If characterization, you need to measure several equal samples in order to detect differents sample's artefacts!

Thank you all,

All the answers made the topic crystal clear to me

thank you researchgate for this oppurtinity to share