Transfer-Function Approach to Substrate-Enhanced Diffraction Tomography

Abstract

Forward and backward scattering provide complementary volumetric and interfacial information, yet conventional three-dimensional (3D) imaging typically accesses only one. In this Letter, we present a substrate-enhanced diffraction tomography approach that simultaneously recovers both channels under multi-angle epi-illumination.This geometry captures one forward- and two backward-scattering bands in axially symmetric Fourier regions, where their complementary coverage enables phase-absorption separation in a non-Hermitian spectrum. Explicit 3D transfer functions are derived for both channels, and an axial Kramers-Kronig relation is established to incorporate substrate-induced boundary conditions in a unified framework. Our results establish a label-free, high-resolution 3D imaging modality that surpasses the limits of existing methods.

Figures (5)

• Figure 1: (a) Passbands in transmission and substrate-enhanced epi configurations; the latter simultaneously records FS and BS, expanding the accessible spectrum threefold. (b) Angle-diverse illumination with varying ${\mathbf{k}}_{\parallel,\mathrm{in}}$ enables retrieval of 3D FS and BS. (c) Substrate-enhanced epi-illumination schematic, the incident field and its substrate reflection act as dual illuminations generating FS and BS. (d) Representative intensity patterns measured under different illumination angles. (e) Linear scattering model expressed via the 3D transfer function.
• Figure 2: (a) Synthetic object above a mirror substrate. The mirror acts as a step function, enforcing causality in the scattering spectra. (b) Real and imaginary parts of the normalized object spectrum at $\mathbf{k}_{\parallel}=(0,0)$; yellow points show $\mathscr{H}_z[\mathrm{Re}(\widetilde{\Delta\epsilon})]$ matching $\mathrm{Im}(\widetilde{\Delta\epsilon})$, demonstrating the axial KK relation. (c) $k_x$–$k_z$ profile of the 3D transfer function for a 0.95 NA under NA-matched incidence, $k_m = \epsilon^{1/2}_mk_0$. (d) NRMSE between the transfer function and MBS models as a function of $\Delta \epsilon$ and volume fraction; insets show representative computed patterns at increasing $\Delta \epsilon$.
• Figure 3: Simulation validation of the reconstruction framework. (a) Ground truth of the synthetic object, with the lower half purely phase and the upper half purely absorptive. (b) $k_x$–$k_z$ profile of the recovered $\widetilde{\Delta\epsilon}$ from reflection-only geometry. (c) $x$–$z$ cross-sections of the reconstructed $\Delta\epsilon_{\mathrm{BS}}$ from reflection-only geometry. (d) $k_x$–$k_z$ profile of the recovered $\widetilde{\Delta\epsilon}$ from substrate-enhanced geometry; right panel: cross-section at $k_x=0.5k_m$ showing $\mathrm{Re}\,\widetilde{\Delta\epsilon}$ and $\mathrm{Im}\,\widetilde{\Delta\epsilon}$ satisfy the axial KK relation. (e) $x$–$z$ cross-sections of the real and imaginary parts of $\Delta\epsilon_{\mathrm{FS}}$ obtained from the FS band. (f) $x$–$z$ cross-sections of the real and imaginary parts of $\Delta\epsilon_{\mathrm{BS}}$ and $|\Delta\epsilon_{\text{BS, sb}}|$ obtained from the combined BS bands and the single BS band, respectively. $\Delta\epsilon_{\mathrm{FS}}$ reveals smooth volumetric features with blurred interfaces, while $\Delta\epsilon_{\mathrm{BS}}$ complements FS by recovering lateral interfaces, with their real and imaginary parts corresponding to the decoupled phase and absorption, respectively. Figures are clipped to positive values for presentation.
• Figure 4: Experimental demonstration with a fixed C. elegans on a mirror. (a) Schematic of the substrate-enhanced epi-illumination LED microscope; inset: LED array and representative measured scattering patterns. (b) and (c) Real and imaginary part of $x$–$y$ cross-sections of $\Delta\epsilon_{\mathrm{FS}}$ and $\Delta\epsilon_{\mathrm{BS}}$ at $z=-23.0~\mu$m (surface) and $z=-3.5~\mu$m (interior). (d) Depth-dependent feature contrast of $\Delta\epsilon_{\mathrm{FS}}$ and $\Delta\epsilon_{\mathrm{BS}}$, showing that the reconstruction is confined to the upper half-space. Inset, zoomed views at five depths.
• Figure 5: Experimental demonstration with a cluster of Chlamydomonas on a mirror. (a) Measured scattering patterns at 515 nm (left) and 632 nm (right). (b) Representative real and imaginary parts of $x$–$y$ cross-sections of $\Delta\epsilon$, with reconstruction at 515 nm in bottom-left panels and that at 632 nm in top-right panels. Insets, zoomed views for detailed comparison. (c) $z$-stack profiles of the framed regions in (b).