Light scatters from travelling compressions
A material that looks still contains moving molecules. Thermal motion creates small fluctuations of density, with slightly compressed and slightly expanded regions that can travel as sound waves. Their changing optical properties scatter a little incident light. An external tap is unnecessary: thermal fluctuations already supply acoustic motion.
Energy exchange with that motion shifts the scattered photons. The lower-frequency signal is called the Stokes peak; its higher-frequency partner is the anti-Stokes peak. Their distance from the laser gives the frequency of the selected acoustic fluctuation. The width of each peak also carries information about how the acoustic motion decays.
Sources: [1]
Collection direction selects an acoustic scale
The incoming and outgoing directions determine which sound waves contribute. Nearly forward collection transfers little momentum. Collection back towards the source transfers the most. At the same sound speed, different acoustic wavelengths have different frequencies, so changing the collection angle moves the peaks.
Refractive index enters the same selection. At a fixed vacuum wavelength, light has a shorter wavelength inside a higher-index medium. The same backscatter arrangement then selects a shorter acoustic wavelength. Its frequency rises even if sound speed stays fixed. A map of peak positions can therefore mix optical and mechanical changes.
A modulus needs more information than a peak position
With refractive index and scattering angle known, a frequency shift gives sound speed. Density is then needed to obtain a longitudinal storage modulus: a measure of elastic energy stored in rapid compression along the propagation direction. Two materials with the same sound speed can have different moduli if their densities differ.
The usual weak-damping relation writes the longitudinal storage modulus as . Here is density, is refractive index, is vacuum wavelength, is the angle between propagation directions inside the sample, and is the fitted frequency shift. The conversion depends on all these inputs.
Consider two model regions with identical sound speed and density. Suppose the second has a refractive index ten percent larger. The same backscatter geometry then gives a ten percent larger frequency shift. If the conversion continues to use the first region’s index, it reports a modulus twenty-one percent larger, because frequency is squared. In this example the acoustic speed has stayed fixed throughout. Measuring the index independently would bring both converted moduli back to the same value.
The result describes a longitudinal response at gigahertz acoustic frequencies. A Young modulus obtained by slowly stretching a sample involves another deformation and another timescale. A viscoelastic material can relax during slow loading and respond differently to rapid compression. Water content can also change density, index and acoustic response together. The interpretation starts with identifying the response being measured.
The instrument contributes its own width
Send a very narrow, known optical sideband into the spectrometer. Its measured peak still has a finite width: the instrument response. A sample measurement combines that response with the material line. Treating the full observed width as acoustic decay would include the spectrometer in the material result.
When both profiles are Lorentzian, with a sharp centre and long tails, their convolution has a full width at half maximum equal to the sum of the two widths. A separately measured response can then be subtracted. Other profiles need their corresponding convolution. If the remaining difference is comparable to counting fluctuations, the original spectrum can be more useful than a precise-looking narrow linewidth.
Sources: [2]
A finite focus can collect two responses
A focal spot has a finite size. Across a material boundary it collects two scattering spectra. Adequate resolution may reveal separate peaks; close responses may instead resemble one broad line. Moving the focus through the boundary and retaining each spectrum with its background lets measured spectra from either side be compared with the mixed signal. A single stiffness need not describe that mixed volume. Resolved responses can be reported separately, together with their photon contributions.