Conception of a coherent imager allowing the visualization of optically diffuse layered material.
This is one of the most recent themes of research in our laboratory. After obtaining very promising preliminary results, the CNRS commissions 4 and 8 and the SPM directors encouraged us to participate actively in setting up a group, originally informal, but later supported by the GDR of Waves and Random Media .
To explain the important points of our work, let us recall that, in order to scan structures which are "drowned" in diffuse media, three approaches may be used. These approaches derive their designations from three classes of photon propagation in the diffuse media from source to detector :
- Balistic (also called coherent) photons which undergo no collisions with the diffusing agents,
- "Snakelike" photons, which are deviated only slightly from their direct optical path based on Fermat's principle, and which may be selected temporaly (these are the first arrivals after a short impuse).
- Photons which trace a complex trajectory described by the diffustion equation which we analyze by inversion to deduce structural details.
We have dedicated our preliminary efforts on the first method by using a 2 wave interferometer (the reference wave allows us to choose either coherent or incoherent photons and eliminate those with stochastically varying phase)
Two setups have been constructed with the principal characteristic of operating with multiple detectors (photodiode arrays or CCD camera), optical path modulation and multiplexed synchronous detection to optimize dynamic detection :
- A Mach-Zehnder type interferometer with a laser source.
- An interferometric microscope with non-coherent sources (LEDs, white light source).
This first setup allowed us to follow interference phenomena (fringes) over a dozen orders of magnitude. For weak diffustion, Mie theory was verified and this model with one particle remains valid. At greater concentrations, two other effects may be produced; multiple diffusion and coherent reconstruction of light diffused in the propagation direction. We have tried to use this setup with biological materials (muscle tissue), but different spatial scales present in these samples create a wave front distortion which leaves only the speckle observable.
To overcome this limitation, we have fixed an home made, immersed interferometric objective with weak magnification onto the microscope.This system allowes us to obtain excellent topographic images of phase objects with resolution on the order of 2 microns across a few millimeters of diffuse media (corresponding to about 10 mean free pathlengths).
Use of a weakly coherent source allows the depth decoupling of objects easily enough, even if the optical quality isn't particularly good.
Projects :
This work will be continued, and applied to the study of practical problems (going beyond "phantoms") in which local optical properties (absorption, diffusion) will have a spatial distribution that we hope to "inverse" based on experimental results. Modelization will take up the support role for instrumentation which has dominated the project to date. The characterization problem for paint (cooperation with St. Gobain, the Musée du Louvre) is being studied, and the basis of a collaboration with biologists at ESPCI and the Institut Curie on brain tissue will be one of our first goals. To put these results in perspective, the modern confocal microscope does not allow exploration at depths greater than a millimeter.
Finally, we wish to attain serpentile photons by high frequency modulation methods similar to those that we have used for quite some time for the immaging of "thermal waves". The diffusion equation solutions give us "light wave density" values, revealing obstacles and discontinuities which begin to be modeled in a realistic fasion at least for simple geometries. back Contact : Prof. A. Claude BOCCARA or email