Fig. 1. Proposed isoFLUX setup. The oblique excitation patterns are generated by the interference of two beams from upper and lower objective lenses, respectively. EOM, electro-optic modulator; M, mirror; L, lens; OBJ, objective lens; DM, deformable mirror; BS, beam splitter.

Fig. 2. Localization precision of isoFLUX versus 3D-SMLM. (a1), (a2) Normalized PSFs (astigmatism and saddle-point) with different depths used in CRLB calculations; scale bar, 1 µm. (b1), (b2) x, y, and z localization precision of isoFLUX and 3D-SMLM from numerical CRLB. (c1), (c2) x, y, and z localization improvement of isoFLUX over 3D-SMLM. (d1), (d2) Line structure reconstructions of isoFLUX (red) and 3D-SMLM (yellow) at different depths under Monte Carlo simulations. The line widths of isoFLUX and 3D-SMLM are fitted by a Gaussian distribution; scale bar, 50 nm. (e1), (e2) Scatterplot of isoFLUX (red) and 3D-SMLM (yellow) at different depths. The standard deviation of the isoFLUX and 3D-SMLM is given aside (isoFLUX: white value; 3D-SMLM: orange value).

Fig. 3. Nuclear pore complex simulation results. (a), (b) Color-coded (depth) 3D images resulting from 3D-SMILE and 3D-SMLM, respectively. (a1), (a2) Zoomed-in views of the white box areas in (a). (b1), (b2) Zoomed-in views of the white box areas in (b). (a3), (b3) Side view cross sections of the white dotted areas shown in (a) and (b). Scale bars: 1 µm for (a) and (b), 500 nm for (a3) and (b3), 50 nm for zoomed-in areas in (a3) and (b3), and 100 nm for others.

Fig. 4. Microtubule simulations. (a), (b) Color-coded (depth) 3D images resulting from isoFLUX and 3D-SMLM, respectively. (a1)–(a3) Side view cross sections along the lines shown in (a). (b1)–(b3) Side view cross sections along the lines shown in (b). Scale bars: 1 µm for (a) and (b), and 20 nm for others.