= N. Everall
{c}
\b[N. Everall] coauthored the 2005 demonstration of <spatially offset Raman spectroscopy>. By changing collection position, the experiment changed the relative surface and subsurface contributions to a measured <Raman spectrum>.<Subsurface Probing in Diffusely Scattering Media Using Spatially Offset Raman Spectroscopy by P. Matousek and others>[exact byline and Abstract]{ref}
The result highlights an experimental distinction between removing an unwanted mean signal and removing its fluctuations. In an original idealized example, suppose a measured spectral channel contains 1000 expected surface counts and 100 expected buried-layer counts. Under independent <Poisson photon counting>, its <variance> is $1100$. Even if the surface mean of 1000 were known exactly and subtracted, the remaining estimate would still have <standard deviation> $\sqrt{1100}\approx33.2$ counts, not $\sqrt{100}=10$ counts.
A surface estimate obtained from a second noisy spectrum adds further uncertainty. Optical collection that changes the mixture before detection can therefore offer something different from subtracting a large background afterward. The hypothetical counts explain a measurement principle; they are not experimental count values attributed to Everall or the 2005 paper.
Back to article page
