Through-container Raman on powders, blends and turbid liquids

Most unsuccessful Raman measurements fail for reasons that have little to do with Raman physics. The scattering cross-section is small but well understood. The difficulty is getting light into and out of a sample that was not designed to be measured.

Powders scatter. Blends are heterogeneous at the scale the spot size samples. Turbid liquids diffuse the excitation before it reaches the material of interest. Many real samples are also inside a container that, for good reasons, should not be opened: a sealed vial, a sack, or an opaque bottle on an incoming-goods dock.

The sampling problem

A conventional dispersive Raman instrument works with a tightly focused spot. That is well suited to a homogeneous, transparent, accessible sample, and it produces three predictable problems elsewhere:

• Sub-sampling. A 50 µm spot on a blended powder measures 50 µm of that blend rather than the blend.
• Fluorescence interference. A small illuminated volume near the surface is where contamination and packaging fluorescence are concentrated.
• No path through a barrier. Focusing at the surface of an opaque container measures the container.

The interferometric approach

An interferometric instrument collects light over a much larger étendue than a narrow imaged slit. In practice this allows illumination and collection over a wide area, which addresses all three problems at once: the measurement averages over a representative volume rather than a point, it collects diffusely scattered light that has traveled through the sample, and it can work through a barrier that scatters rather than absorbs.

The trade-off should be stated plainly. Some spectral resolution is given up compared with a high-dispersion dispersive instrument, and the data reduction is a transform rather than a direct read. For identification, verification and quantification of major components, the trade is worth making. For resolving closely spaced bands in a research setting, it may not be.

Applications suited to the method

• Incoming raw-material verification without breaking a seal
• Blend uniformity, where sub-sampling must be avoided
• Turbid or highly scattering liquids
• Any measurement where opening the container changes what is measured