The proof-of-concept demonstration of holographic flow cytometry presented here can be improved upon in several respects. The substrate beads' size and composition, selected here for convenience, can be optimized for sensitivity. Analysis of calculated scattering patterns suggests that somewhat smaller spheres made of a lower-index material such as silica would offer greater sensitivity to the presence of molecular-scale coatings. This added sensitivity would be useful for detecting smaller molecules, and could be sufficient to seek out nonuniformity in individual spheres' surface coatings.
Comparison with similar GPU-based applications suggests that processing time, which presently stretches to several hours, can be reduced by another order of magnitude on existing computer hardware through rigorous software optimization. In that case, population-based molecular binding assays of the kind we have presented could be completed in several minutes.
Greater sensitivity and speed also could be attained through an optimized choice of laser wavelength. Simultaneous measurements in two or more wavelengths might even enable detection of molecular coatings on individual spheres while also providing spectroscopic information on the coatings' composition.
Even in its present form, the method we have presented here offers precision, simplicity, generality and speed for colloidal tracking and characterization. Consequently, holographic analysis should prove useful in other application areas. For example, high-resolution single-particle characterization is superior to bulk light-scattering for probing the properties of mixed colloidal samples because it does not rely on models for those properties' distributions. Holographic characterization therefore should be a useful adjunct for colloidal synthesis, and is rapid enough to be useful for process control. Nanometer-resolution three-dimensional holographic tracking data already has proved useful for microrheology (9) and research in statistical physics (8). The method's comparative simplicity and use of off-the-shelf components should encourage rapid adoption in areas that previously have been dominated by conventional light microscopy.
This work was supported by the National Science Foundation through Grant Number DMR-0606415 and by the Keck Foundation. B.S. acknowledges support of the Kessler Family Foundation.