The principle answer to your questions is, yes you can do Raman spectroscopy with any excitation wavelength. That is the whole idea/principle of inelastic scattering. Therefore, people give Raman shift, as independent value of the excitation wavelength. After all, it just gives you the frequency/wavenumber/energy of the inelastic scattering partner of your photon. However, there are a lot of things to consider:

a) (Spontaneous) Raman scattering is dipole scattering, therefore it scales with the 4th power of the excitation frequency. Therefore, using a shorter wavelength, may be desired. This very general rule, also gives you an easy idea, why Raman cross-sections get very very small soon, if going to a very long excitation wavelength (Mid IR or IR). One a side note: A shorter wavelength might also give you a better spatial resolution, if you are doing µ-Raman (=Raman spectroscopy in a microscope system with highly focused light). But spatial resolution just scales linearily with wavelength, it is only a very slight effect.

b) Using a too short wavelenght may result in absorption by the electronic system of your sample. This has two consequences. i) Light may be reemitted from your sample in terms of photoluminescence, which can be much brighter than your Raman signal, making it difficult to actually distinguish your Raman light in your spectrum. After all (spontaneous) Raman scattering only happens to something like 1 in 10^6-10^8 photons. Therefore, people often like to use near IR sources, as many molecules (or semiconductors, like Si, GaAs) are non absorbing for wavelength longer than ~0.8-1 µm. ii) If your light is absorbed at the surface, you will not be able to do 3D analysis and measure spectra below the sample surface, e.g. in confocal µ-Raman or Raman Imaging applications.

c) Choosing specific wavelength, may have benefits or additional effects: For example, Resonant Raman scattering. If your excitation wavelength coincides with a real energy transition, usually an electronic absorption, in your system, this can make the Raman effect very efficient and you will see an order or two of magnitude stronger Raman signal and sometimes even overtones of your Raman signal, i.e. two or three times scattering with the same phonon or vibration. This is for example seen in Raman scattering of certain semiconductors (e.g. in II-VI like CdSe or something like this, if I remember correctly.). I am pretty sure, there are examples from spectroscopy of molecules, as well. Resonant Raman scattering can be understood in a quantum mechanical picture of the Raman effect. In this picture Raman scattering involves the transition of you energy to a virtual, short-lived, intermediate level, before your new, Raman-shifted photon is emitted. If this virtual level conincides with a real level (and transitions are allowed), the Raman effect can become much more efficient.

d) For certain materials, often also having resonance effects, shifting the wavelenght may also shift your Raman peaks. One example coming to my mind is the defect scattering in graphene. Here, certain defect peaks shift with the excitation wavelength, as also electronic transitions (with energies and k-vectors) are involved. Therefore, sometimes people compare and publish spectra measured with different excitation wavelenght, to prove (or disprove) that such effects are present (or not).

e) I am pretty sure, there are many more specialized examples.

I hope, this gave you a better overview,

Regards,

Michael

The principle answer to your questions is, yes you can do Raman spectroscopy with any excitation wavelength. That is the whole idea/principle of inelastic scattering. Therefore, people give Raman shift, as independent value of the excitation wavelength. After all, it just gives you the frequency/wavenumber/energy of the inelastic scattering partner of your photon. However, there are a lot of things to consider:

a) (Spontaneous) Raman scattering is dipole scattering, therefore it scales with the 4th power of the excitation frequency. Therefore, using a shorter wavelength, may be desired. This very general rule, also gives you an easy idea, why Raman cross-sections get very very small soon, if going to a very long excitation wavelength (Mid IR or IR). One a side note: A shorter wavelength might also give you a better spatial resolution, if you are doing µ-Raman (=Raman spectroscopy in a microscope system with highly focused light). But spatial resolution just scales linearily with wavelength, it is only a very slight effect.

b) Using a too short wavelenght may result in absorption by the electronic system of your sample. This has two consequences. i) Light may be reemitted from your sample in terms of photoluminescence, which can be much brighter than your Raman signal, making it difficult to actually distinguish your Raman light in your spectrum. After all (spontaneous) Raman scattering only happens to something like 1 in 10^6-10^8 photons. Therefore, people often like to use near IR sources, as many molecules (or semiconductors, like Si, GaAs) are non absorbing for wavelength longer than ~0.8-1 µm. ii) If your light is absorbed at the surface, you will not be able to do 3D analysis and measure spectra below the sample surface, e.g. in confocal µ-Raman or Raman Imaging applications.

c) Choosing specific wavelength, may have benefits or additional effects: For example, Resonant Raman scattering. If your excitation wavelength coincides with a real energy transition, usually an electronic absorption, in your system, this can make the Raman effect very efficient and you will see an order or two of magnitude stronger Raman signal and sometimes even overtones of your Raman signal, i.e. two or three times scattering with the same phonon or vibration. This is for example seen in Raman scattering of certain semiconductors (e.g. in II-VI like CdSe or something like this, if I remember correctly.). I am pretty sure, there are examples from spectroscopy of molecules, as well. Resonant Raman scattering can be understood in a quantum mechanical picture of the Raman effect. In this picture Raman scattering involves the transition of you energy to a virtual, short-lived, intermediate level, before your new, Raman-shifted photon is emitted. If this virtual level conincides with a real level (and transitions are allowed), the Raman effect can become much more efficient.

d) For certain materials, often also having resonance effects, shifting the wavelenght may also shift your Raman peaks. One example coming to my mind is the defect scattering in graphene. Here, certain defect peaks shift with the excitation wavelength, as also electronic transitions (with energies and k-vectors) are involved. Therefore, sometimes people compare and publish spectra measured with different excitation wavelenght, to prove (or disprove) that such effects are present (or not).

e) I am pretty sure, there are many more specialized examples.

I hope, this gave you a better overview,

Regards,

Michael

Mr. Hekmati,

Very important!

Thanks Aurelio for sharing the link

I read the paper. It was very good.

Thanks Michael,

Your explanation is very helpful and instructive.

The wavelength of the laser is chosen so that there is no luminescence and heating of the sample

Dear Reza Hekmati! I do agree with the answer of Michael Rüsing.

And practically, answering your question, I would add that as far as

the resonant Raman scattering is not observed frequently, you are

right. You have to avoid laser wavelength that would excite your

sample intense luminescence. But in many cases the luminescence

(if it is not too strong) produce a pedestal at the spectrum, on top

of which you still can see the Raman line. So you may first try the

laser you have and then decide.

Dear Reza, The below link gives a good insight about selecting the correct laser wavelength for your sample.

Selecting the Correct Laser Wavelength for Raman Material Identification. Link below

https://www.azom.com/article.aspx?ArticleID=11871

Thanks.

Dear Michail,

Thanks

Dear Md ABDULLAH,

Thank you for sharing the link