Imagine that you are focusing the raman laser right at the center of the vial sample with water (or any liquid) using 532nm laser.

enter image description here enter image description here (from "Photometric Standards for Raman Spectroscopy)

For raman 180 degree backscattering sample geometry, can the inelastic scattering photons upon travel to the spectrometer becomes elastically scattered on way to it by the thick liquid sample such that it is scattered outside the area?

If yes, does it mean it is better not to focus so deep into the sample so the return path would be less (that is focusing the raman laser just after the glass on the liquid sample)? The side effect of this is getting some raman noises from the glass itself.


2 Answers 2


Yes, of course the Raman scattering can suffer from additional scattering on its way through the sample, just like it will generally be attenuated on its way to the detector by any element in its path (mirrors, lenses, glass surfaces, etc.).

If scattering is a problem or not very much depends on your sample. The ideal sample will not scatter your excitation laser or Raman scattering much, like for instance a clear solution, and then I would say the influence on the measurement is negligable. If you have something opaque, such as an emulsion, additional scattering will be relevant and you would want to measure near the sample surface.

If you have problems with a glass-affected background you could switch to quartz glass or use a confocal Raman setup, if available.

And also depending on your sample, absorption might be something you should consider as well.

By the way: I would probably rather go for a planar glass surface in order to avoid refraction as it occurs on a curved glass surface.

Edit due to comments: Pure water does neither absorb nor scatter light with a wavelength around 532 nm much. If it would, you would be able to see this with your eyes. Wikipedia claims that a fraction of 10-7 of the incoming light is scattered inelastically. Although this should be only used as a rule of thumb, this tells you that a loss of Raman intensity due to "secondary" Raman scattering is not a practical problem you would usually be able to observe. Also, I would not be worried about elastic scattering although it is much more relevant than inelastic scattering (Wikipedia claims 0.1% to 0.01%).

I do not know your Raman setup in detail. But if you have anything like a confocal aperture (pinhole, slit...) on the path to the detector, refraction on the container/water surface will definitely reduce your signal intensity with increasing penetration depth, especially if you do not measure with an immersion objective lens. This is because your focus is severely distorted especially in length when focussing deeper into the sample, it will become much longer. Your confocal aperture then blocks part of your signal because it originates from a part of the sample which is far from the focal plane. This is descibed in detail for example in this paper, and also in this paper. However, the papers are behind a paywall.

  • $\begingroup$ What puzzled me is water is affected too. I used a plastic bottled water commonly used. When I sampled near the container surface, the water OH peak counts per frame at 3400 cm-1 is higher (about 14,000 counts) than if the focus is deep inside the plastic bottled water (only 10,000 counts). So if water can be raman scattered by the 532nm system, it can suffer rescattering on the way back to the spectrometer too? $\endgroup$
    – Jtl
    Oct 20, 2019 at 12:17
  • $\begingroup$ To continue.I think what I wanted ultimately to know is in the case of water. Is it scattering or absorption that can change the counts from 14,000 to 10,000 using 532nm laser? What do you think based on your experience with water raman? $\endgroup$
    – Jtl
    Oct 20, 2019 at 12:57
  • 1
    $\begingroup$ How deep is deep inside the bottle? $\endgroup$ Oct 20, 2019 at 13:16
  • $\begingroup$ laser focused 20mm into the bottled water versus just 3mm. So is it scattering or absorption that mostly lower the counts. What do you think? $\endgroup$
    – Jtl
    Oct 20, 2019 at 13:34
  • 1
    $\begingroup$ "Wikipedia claims that a fraction of $10^{-6}$ of the incoming light is scattered inelastically". We need here the fraction that is scattered elastically since the discussed mechanism of loss is that a Raman scattered photon is then undergoing a second scattering process. That could in principle be elastic or inelastic, but elastic is several orders of magnitude more likely, so p (further scattering) ≈ p (further elastic scattering). It is still only a tiny fraction of of the Raman scattered intensity. $\endgroup$ Oct 21, 2019 at 15:05

A few points in addition to @Alexander's answer:

  • Total scattering (including secondary) losses depend on the optical pathlength (cuvette thickness).

    • for 10 mm clear water, I'd expect this to be $\ll 1 \%$, so it doesn't explain the 30 % loss you encounter.

    • Yes, secondary scattering can sometimes be noticeable in Raman setups. In clear media it plays a role only for long interaction lengths (e.g. collection fibers that have the Rayleigh filter only after the fiber may generate background Raman signal of the fiber)

    • Practical plausibility check: If your laser intensity (via laser intensity meter) after a flat of same thickness as your cylindrical vial isn't noticeably lower than before, then there isn't much of an intensity loss due to any type of scattering over the whole thickness of the cuvette, and in conseqence also not for the Raman scattered photons travelling towards your detector.

  • Even for geometries with low confocality, the optical geometry with the round cuvette may be the problem here: if the round cuvette is positioned (slightly) off the optical axis, you get higher reflection losses at the surface (the same is true if a flat cuvette is tilted).

    • A symptom of this should be that the observed intensity varies a lot when taking out the cuvette and putting it in again.
    • You can again check laser intensity before and after the cuvette: as long as you still get the whole beam onto the sensor area of the laser power meter, higher losses than though a similar flat cuvette are reflection losses (though both entering and leaving the cuvette)
  • In addition, if you use a high NA objective, that has a wide angle collection cone.

    • This is typically desired because you collect Raman signal over a wide solid angle and thus get better intensity.
    • But unless you manage to get your cuvette positioned so that the center of the cuvette is where the focus point in air would be (i.e. collected light leaves the cuvette at 90° to the cuvette surface), you may get substantial losses due to reflection:
      reflection losses at glass to air transition

      (see Wikipedia on Fresnel equations)

    • Note that these reflection losses are polarization dependent, so may affect different bands differently depening on their depolarization ratio.

    • Again, an immersion may help.

In consequence, you need to optimize depth into the cuvette (and possibly cuvette diameter), lateral position (centered) and the angle of the cuvette (wrt. polarization) in order to get maximum intensity.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.