A new way of recording three-dimensional (3-D) information—called spatiotemporal digital holography—should give researchers the option of recording data first and deciding exactly what they need to extract from it later. Developed by researchers Guy Indebetouw and Prapong Klysubun at Virginia Tech (Blacksburg, VA), the technique uses light that is only spectrally (rather than spatially) coherent. This, they say, removes two of the problems of conventional holographic methods: needing a detector with high spatial bandwidth, and dealing with the artifacts of coherence such as speckle noise. The team has demonstrated how the relatively few images acquired can be used to produce phase maps of the reconstructed field, differential phase contrast images, and real and imaginary parts of the complex field. All were achieved through post-processing.
The basic optical system is an interferometer, the imaging arm of which has a large defocus. When the light traveling along this arm scatters from the specimen, the output from the optical system is essentially a plane wave. The reference beam, on the other hand, is focused through a pinhole, producing a spherical wave. When these two beams interfere, say researchers, the result is a Fresnel pattern that encodes the 3-D position of the original scattering object.
What makes the Virginia Tech system different, however, is the way that it exploits the speed of its CCD sensor to allow information to be captured in time as well as space.1, 2 Instead of using the precise phase-stepping techniques common in interferometry, the system is set up to allow many different images to be recorded one after another, each with a slightly different phase difference between the object and reference beam. This is achieved by having the path difference modulated by a piezo-electrically positioned mirror controlled by a sawtooth signal. Thus, what the holographic system lacks in spatial bandwidth (number of pixels in the detector) can be made up for by the temporal bandwidth (number of frames, each containing new information).
Once sufficient frames have been recorded—theoretical analysis shows that this number can be as low as four and still satisfy the Nyquist sampling theorem—the information from the various frames can be combined digitally. Reconstructions can then be tailored to suit the needs of the researcher. For instance, researchers recorded several frames from an unstained blood smear, from which they were able to reconstruct amplitude, phase, and quantitative phase maps.
More complicated functions can be achieved by, for instance, using phase distributions taken from different spatial positions and subtracting them from each other to produce a differential phase contrast image. Real and imaginary parts of the reconstructed hologram can also be pulled out and used, and the information extracted can be tailored even more by digitally implementing various reconstruction functions.
- G. Indebutouw and P. Klysubun, J. Opt. Soc. Am. A 18 (2), 319 (February 2001).
- P. Klysubun and G. Indebutouw, J. Opt. Soc. Am. A 18 (2), 326 (February 2001).