(DMDs) to maintain consistency across the projected field. Although, in principle, one can directly use a Gaussian laser beam to
efficiently illuminate phase holograms, this can be problematic
when using liquid crystal-based devices (e.g., LCoS) for high power
applications due to the intensity hotspot at the center of the beam
and the broadened point spread function. In this case, uniform
illumination presents a practical solution by spreading the incident power to achieve higher power throughput within the
power limits of the device.
The most common method for illuminating beam shaping devices – expanding and truncating the laser beam – wastes photons
and achieves uniform illumination by sacrificing light efficiency.
Yet, many beam shaping applications demand high efficiency. In
the work by Kato et al. on multi-spot parallel microfabrication [8],
they needed to amplify the laser source to address a fixed microlens array. In this case the available laser power limits the extent to which processes can be parallelized. Since energy is distributed among the focal spots, increasing the number of focal
spots creates spots with lower intensities. This problem is compounded for applications based on two-photon excitation, which
depend quadratically on intensity [9]. These power considerations may be mitigated, if one can afford higher power sources
(although commercial availability may be difficult for some wavelengths). Nevertheless, efficient energy usage is always desirable
and should always be encouraged in any optical engineering
design. More recently, the feasibility of a 2000-fold parallelized
dual color STED fluorescence nanoscopy has been reported [10].
The lateral resolution of STED nanoscopy is dependent on intensity
and therefore such massive parallelization would require high
intensity input pulse. In this case, “the STED pulse energy is a