Figure 2
Illumination around large refractive objects and optimization of beam parameters (DSLM vs. mDSLM). (a) Simulation and corresponding experimental image of a circular Gaussian beam (NAx = NAy = 0.06) propagating around a large glass sphere (diameter d = 20 μm, nsphere = 1.59) embedded within a fluorescent gel (ngel = 1.46). The sphere is positioned at a depth of zsphere = 125 μm at an offset of Δy = 2 μm from the optical axis of the pencil beam. The pencil beam focus is located at a depth of zfocus = 350 μm. For a circular Gaussian beam (DSLM), the intensity at the beam focus is reduced by >75% relative to an unobstructed beam, as illustrated by the overlaid line profiles. (b) Simulation and corresponding experimental image of an elliptical Gaussian beam (NAx = 0.06, NAy = 0.18) propagating through an identical fluorescent gel and glass sphere. In contrast to the circular Gaussian beam used in DSLM, the increased angular diversity in the y direction enables the elliptical Gaussian beam (used in mDSLM) to experience only a ~10% reduction in intensity relative to an unobstructed beam. Simulated and experimental beam-scanned images, with DSLM and mDSLM, are shown in (c). Simulation results are plotted in (d) for the dependence of the beam-focus intensity (behind the glass sphere) as a function of NA (NAx = 0.03–0.24, gray lines, and NAy = 0–0.24, horizontal axis, in increments of 0.03). Panel (e) provides simulation results for the dependence of the fluorescence signal at the field edges (z = ±zR,x), as a function of NAx and NAy (in the absence of the glass sphere). In both (d,e), the black solid line indicates the value of NAx (0.06) used experimentally in the majority of our studies, in which the experimentally used values of NAy are indicated as yellow points (0.06 and 0.18 for DSLM and mDSLM, respectively). Illustrations are shown above the horizontal graph axes in (d,e) to depict the reduction in shadowing artifact as a function of NAy (d), but an increasing roll off in signal at the field edges (e).
