Indeed Low-frequency Raman spectroscopy was not trivial and tests the limits of filters that are based on thin film technology. Lately, it has become increasingly common to use low-frequency Raman to detect crystalline polymorphs of active pharmaceutical ingredients in the pharmaceutical industry due to developments in filter technology.
The classical way to get below a 100 cm-1 is using a triple spectrograph in subtractive mode, as a few people have already suggested. The advantage of using this method is tunability, it might be used with different excitation wavelengths. However, a major drawback is that triple spectrometers have notoriously low throughput. This can prohibit spectra from being collected in a reasonable amount of acquisition time. This is often limiting, especially in industrial or process environments.
One way to circumvent the throughput problem is using Ondax's THz-Raman systems. These systems allow measurements down to 10 cm-1 using an ultra-stabilized laser, patented VHG gratings and carefully designed optical assembly. The fiber output from the THz-Raman system is coupled to a spectrograph and the spectral resolution is governed by the choice of spectrograph. There is a much higher throughput using THz-Raman instrumentation and other 'soft advantages' when compared to a triple spectrograph. Attached are links to a few publications that utilize THz-Raman systems for low-frequency Raman spectroscopy.
In summary, there are pros and cons to either path and researchers should carefully consider and weigh them according to their application.
Article Direct High-Resolution Imaging of Crystalline Components in ...
Article Investigating Crystallinity Using Low Frequency Raman Spectr...
Can you share the Raman spectrum you see, and maybe the material you have as the thin film? The answer could be one of many possibilities, and the data may help to narrow down to just a few possibilities.
When I worked with Raman measurements, I put the spectrometer grid in 50 cm-1 to prevent the CCD device was damaged. So I think you can get Raman measurements below 100 cm-1.
It is definitely possible. However, you need special equipment to get closer to the Rayleigh line. There are dedicated spectrometer, sometimes called Terahertz-Raman Systems with which you can get down as far as 5 cm-1 or even down to 0.1 cm-1. See e.g.
Yes it possible but you should use proper equipment. As usual this is a spectrometers with several grating like Horiba Jobin-Yvon T64000 ( http://www.horiba.com/scientific/products/raman-spectroscopy/raman-spectrometers/triple-raman-spectrometers/details/triple-raman-spectrometers-140/ ) or U1000 ( http://www.horiba.com/fileadmin/uploads/Scientific/Documents/Raman/HSC-T64000_U1000-2013-V1.pdf ).
For single grating spectrometers it is possible with Brag Notch Filter (for example http://www.optigrate.com/BragGrate_Notch.html )
I completely agree with A.S. Krylov. Optigrate is the way to go for good Bragg Gratings dedicated to a grating spectrometer.
On a different topic, we have long preferred not to use slow throughput spectrographs as they are not entirely necessary to achieve desirable S/N contrasts near the Rayleigh peak. For diamnd-anvil cell HP research, we can get down to ~ 10 wavenumbers using fast spectrographs, confocal imaging, and the correct filters. But again, our micro-Raman systems (including micro-FTIR) are all custom designed/built for high-performance.
In my lab I have Horiba Jobin-Yvon T64000 - on this instrument is possible go down up to 3-5 cm-1 at any excitation wavelength without of any filters (Depends on sample). This instrument no so fast in measurements but it specially designed for extra low wavenumbers.
I fully agree with the colleagues. In our lab we have implemented a Bragg notch filter stage (from Ondax and/or Optigrate) into a Renishaw RM1000 Raman spectrometer (single stage) and can easily get down to 5 cm-1 from the laser line both on the Stokes and anti Stokes side.