The 2D approximation quickly breaks down in reflection ptychography Journal: opticajournal Publication type: Research Article

Sander Senhorst 1,*, Stefan Witte 1 and Wim Coene 1,2 · 2026

Read The 2D approximation quickly breaks down in reflection ptychography Journal: opticajournal Publication type: Research Article by Sander Senhorst 1,*, Stefan Witte 1 and Wim Coene 1,2 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-05989.txt.

Introduction

Senhorst, Witte, and Coene test the thin-sample model behind reflection ptychography, an imaging method that rebuilds samples from scattered light. A three-dimensional weak-scattering account sets thickness limits for two-dimensional models, and reflection proves far stricter than transmission. Allowed thickness drops by one to two orders of magnitude in a sample extreme ultraviolet setup, and depth-aware propagation fixes the distortions.

Catalog record

CREATORSander Senhorst 1,*, Stefan Witte 1 and Wim Coene 1,2
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE49999
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.05989

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The 2D approximation quickly breaks down in reflection ptychography Journal: opticajournal Publication type: Research Article Sander Senhorst 1,*, Stefan Witte 1 and Wim Coene 1,2 Abstract Ptychographic reconstructions in reflection geometries are commonly interpreted with the same two-dimensional thin-sample model used in transmission, yet the validity of this approximation has not been established. We develop a three-dimensional weak-scattering description of reflection ptychography and derive explicit thickness criteria for when a two-dimensional model remains accurate. Because the sampled axial spatial frequency range is dominated by the rotation of the Ewald sphere rather than its curvature, reflection geometries impose far stricter thin-sample conditions than transmission geometries. The allowable thickness is reduced by one to two orders of magnitude for a representative extreme ultraviolet geometry, depending on the tolerance for appearance of artifacts. Simulations verify that conventional two-dimensional reconstructions may exhibit the thickness-dependent artifacts as predicted by the theory, with particularly strong distortions near specular Bragg minima. We further show that incorporating the correct depth-dependent propagation into the forward model resolves these distortions and enables recovery of sample thickness. These results establish practical validity limits for two-dimensional reflection ptychography and identify a path toward quantitative depth-sensitive reconstructions at all geometries. 1 Introduction High-resolution imaging of surface structures enables advancement in many areas of research and technology, such as the imaging of patterned nanostructures for the field of lithography. As lithographic functional designs are set to create samples of interest with ever smaller lateral features, stacked on increasingly taller stacks [ 1 ] , ptychography in reflection geometries, specifically at extreme ultraviolet (EUV) wavelengths around 13.5 nm, has emerged as a potential imaging solution that could tackle the needs of metrology and inspection for future lithographic processing nodes. Ptychography is a form of coherent diffractive imaging (CDI), where a coherent probe beam is scanned across a sample, while at each scanning position a diffraction pattern is recorded. If the scanning is performed with sufficient overlap in the illumination of adjacent probe beam positions, the resulting collection of diffraction patterns provides enough redundancy in the data to over-constrain the otherwise ill-posed inverse problem of reconstruction of the sample function from the data. This process, known as lensless imaging, enables the reconstruction of sample images at diffraction-limited lateral resolutions without the requirement of diffraction-limited optical components. Ptychography is of special benefit in application domains where use of diffraction-limited optics is prohibitive due to high cost or difficulty in manufacturing, such as in the domain of imaging using extreme ultraviolet wavelengths. Since many structures of interest are placed on substrates which are optically opaque, either due to thickness or strong absorption, it is often not possible to measure sufficient scattered intensity in the transmission direction, so for such samples one must resort to imaging in a reflection-type geometry. An additional complication for EUV wavelengths is that the reflection coefficients only become significant at near-grazing incidence angles. For this reason, the angle of incidence on the sample is often chosen in the (60^{\circ},85^{\circ}) -degree range [ 5 , 15 ] , leading to the additional complication of diffraction-pattern distortion. The theory of ptychography was originally developed for the transmission geometry [ 14 ] . For reflection geometries, its mathematical validity was never verified, despite many experimental demonstrations of its feasibility [ 5 , 15 , 11 , 22 , 10 , 17 , 13 , 7 , 20 , 2 , 9 , 18 , 19 , 16 ] . In many of the experimental demonstrations of reflection ptychography the reconstruction phase is converted to the height by means of the optical path difference and phase shift upon reflection [ 9 , 18 ] , thus modeling the effect of depth in the sample as a constant phase shift in real space. Such an approach is based on the approximation that considers single incoming and outgoing wavevectors. The bounds of this approximation are however not clearly defined. By its nature, ptychography requires a range of incoming and outgoing wavevectors to generate diversity in the detected intensity, e.g. to create a focused illumination probe. Such a model can therefore only be completely correct if all incoming and outgoing wavevectors feel an approximately constant phase shift due to sample height, which should hold only in the limit of small detection and illumination numerical apertures. Some attempts have been made to model the effect of depth in reflection ptychography [ 4 ] , which models layers as incoherent contributions, requiring explicit filtering of interference terms through spatial separation, and therefore only works for large thicknesses, such as modeling the reflection from the bottom of a subtrate. In this work, we develop a theory that can model the effect of depth with explicit bounds on the maximal permissible thickness, using the same foundational assumption of weak scattering as the original ptychographic theory. Our developed theory predicts that distortions appear in reconstructions that are performed without adequately incorporating depth effects, which we verify in simulation. Finally, we show that more accurately incorporating the effect of depth in the ptychographic model can allow reconstruction of the sample thickness using the propagation geometry alone. 2 The 2D ptychographic model In 2D scanning ptychographic reconstructions assumes the separability of the field \psi emerging from the sample of interest as the product of two parts: one which is independent with respect to the applied shift vector \boldsymbol{R}_{k} at positional index k , and the other which is translated with respect to \boldsymbol{R}_{k} , i.e. \psi_{k}(\boldsymbol{r})=O(\boldsymbol{r}-\boldsymbol{R}_{k})P(\boldsymbol{r}), (1) where in this case we have taken the “object” function O to be the shifting part, and the “probe” function of the illumination P to be the independent part. If the measurement is performed in the Fraunhofer regime, we detect the intensity I_{k} of the Fourier transform of the exit field \psi_{k} , i.e: I_{k}(\boldsymbol{\xi})=|\tilde{\psi}_{k}|^{2}=|\mathcal{F}_{2}\left\{\psi_{k}\right\}|^{2}, (2) where \boldsymbol{\xi} is the reciprocal space coordinate measured by the detector and the tilde is used two-dimensional functions in reciprocal space, with the two-dimensional Fourier transform represented by \mathcal{F}_{2} . Purely two-dimensional samples such as O in Eq. ( 1 ) exist only in theory, yet the preceding analysis is purely two-dimensional in both the real space ( \boldsymbol{r} ) and reciprocal space ( \boldsymbol{\xi} ) coordinates. 3 Transmission and reflection geometries Before continuing with derivations, it is of benefit to already define specific geometries of our interest as this may aid in the visualisations by means of example. These geometries are the transmission and reflection geometries, which we parameterize in terms of • the central illumination wavevector \boldsymbol{k}_{i,0}=k_{0}(\sin\theta_{i},0,\cos\theta_{i}) , • the largest angle \Delta\theta_{i} away from \boldsymbol{k}_{i,0} such that the illumination numerical aperture \text{NA}_{i}=\sin\Delta\theta_{i} , • the central detection wavevector \boldsymbol{k}_{d,0}=k_{0}(\sin\theta_{d},0,\cos\theta_{d}) and • the largest angle \Delta\theta_{d} away from \boldsymbol{k}_{d,0} such that the illumination numerical aperture \text{NA}_{d}=\sin\Delta\theta_{d} , where \theta_{i} and \theta_{d} are the angles with respect to the z-axis. Defining a transmission geometry is easily done by simply take any geometry where \theta_{i}=\theta_{d} . For simplicity we choose \theta_{i}=\theta_{d}=0 such that the the central direction of incidence is aligned with the z-coordinate. The detected part of the exit field is then propagating in the same z-direction as the illumination part. Choosing \theta_{i}=\theta_{d}\neq 0 does change the requirements on the thickness, but only because

Read the complete text