🔭 astrophysics

Sequential and non-sequential Zemax Dynamic Link Libraries for generating image slicer integral field units

This paper introduces sequential and non-sequential Zemax Dynamic Link Libraries that efficiently and accurately model image slicer integral field units within a single file, overcoming the limitations of traditional multi-configuration approaches and successfully replicating the design of the SPECTRE spectrograph.

Original authors: Ellen Lee

Published 2026-04-20

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Ellen Lee

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are an astronomer trying to take a picture of a distant galaxy. But you don't just want a photo; you want to know the chemical makeup of every single star in that galaxy at the same time. To do this, you need a special tool called an Integral Field Unit (IFU), specifically one that uses an Image Slicer.

Think of an image slicer like a very fancy, high-tech slicer of a loaf of bread.

The Problem: The "Loaf of Bread" Puzzle

In a normal telescope, you look at a whole scene at once. But to get a detailed spectrum (a chemical fingerprint) of the whole scene, you usually have to cut the image into thin strips, like slicing a loaf of bread, and feed those strips one by one into a spectrometer.

However, an Image Slicer does something magical: it takes that whole loaf of bread, slices it into 36 (or more) tiny strips, and then rearranges them side-by-side so they all fit into the spectrometer's "mouth" at the exact same time. This allows the telescope to capture spatial and spectral data simultaneously.

The Catch: Designing these slicers in computer software (specifically a program called Zemax, which is the industry standard for designing lenses and mirrors) is a nightmare.

The Old Way: To model a slicer with 36 slices, the old method forced the computer to treat every single slice as a completely separate "file" or "configuration." It was like trying to build a house by designing every single brick in a separate room and then trying to glue them together. It was slow, prone to errors, and the computer would often crash or take hours to calculate a single ray of light.
The Diffraction Issue: Light doesn't just travel in straight lines; it bends and spreads out (diffraction). The old method couldn't easily calculate how light from one slice would spill over onto its neighbor, which is crucial for getting a clear image.

The Solution: The "Magic DLL"

The author of this paper, Ellen Lee, created a Dynamic Link Library (DLL). If Zemax is the kitchen, the DLL is a specialized, pre-made appliance that you plug in.

Instead of building the slicer slice-by-slice in the software, this new tool lets you describe the entire slicer as one single, smart object.

Here is how it works, using some analogies:

1. The "Smart Blueprint" (Sequential Mode)
Imagine you have a blueprint for a complex building. Instead of drawing every single brick individually, you tell the computer: "Here is the shape of the brick, here is the angle, and here is the rule for how they are stacked."

The Innovation: The new DLL acts like a "smart blueprint." You input the rules (how many slices, how they are tilted, how they are curved), and the DLL instantly generates the entire optical path.
The Benefit: It's like switching from drawing a map by hand to using Google Maps. The computer can now trace millions of light rays through the whole slicer in seconds instead of hours. It can also calculate the "spillover" of light (diffraction) between slices, which was previously impossible to do accurately in a single file.

2. The "3D Lego Block" (Non-Sequential Mode)
Sometimes, engineers need to see how stray light bounces around the inside of the instrument (like light leaking through cracks).

The Innovation: The DLL can also build a 3D "Lego block" version of the slicer. It creates a solid, closed object made of tiny triangles (facets).
The Benefit: This allows the software to simulate light hitting the edges and steps between the slices. It's like having a 3D printer that can instantly print the slicer so you can see exactly where light might get trapped or scattered, helping engineers design better shields to block that unwanted light.

Why This Matters

Before this paper, designing these complex instruments was like trying to solve a Rubik's cube while wearing oven mitts—slow, clumsy, and frustrating.

Speed: The new tool is exponentially faster. A design that used to take hours to update now takes seconds.
Accuracy: It can model the entire instrument in one file, meaning no more "gluing" separate files together, which reduces the chance of mistakes.
Real-World Impact: The author tested this on SPECTRE, a real, high-tech instrument being built for the NASA Infrared Telescope Facility. The DLL successfully recreated the design of SPECTRE's 36-slice slicer, proving it works for real-world astronomy.

The Future

The paper suggests that this tool is just the beginning. Because the DLL is so flexible, it can be used to model other complex grids of surfaces, like microlens arrays for 3D medical imaging or digital mirrors for space telescopes.

In short, Ellen Lee has built a universal translator for complex optical slicers. She turned a tedious, error-prone manual process into a streamlined, automated workflow, allowing astronomers to design better instruments faster and see the universe more clearly.

1. Problem Statement

Integral Field Spectrographs (IFSs) utilizing image slicer Integral Field Units (IFUs) are critical in astronomy for simultaneous spatial and spectral mapping. However, modeling these complex optical systems in Zemax OpticStudio presents significant challenges:

Sequential Mode Limitations: Traditional modeling requires splitting the image slicer into multiple "configurations" (one per field channel/slice). This approach is computationally expensive, prevents the analysis of combined diffraction effects (Physical Optics Propagation), and limits the ability to optimize the system or model multiple spectral channels within a single file.
Non-Sequential Mode Limitations: While non-sequential mode allows for complex geometry, it lacks paraxial analysis features and diffraction simulation capabilities. Furthermore, converting a sequential design to non-sequential often involves exporting to CAD, merging surfaces, and re-importing, which introduces coordinate system errors, reduces ray-tracing speed, and makes optimization difficult.
Lack of Standardization: There is no unified tool to generate image slicer prescriptions from basic system requirements, and existing methods struggle to handle the diverse morphologies of modern slicers (e.g., varying curvatures, tilted stacks, and complex pupil re-imaging).

2. Methodology

The author developed a set of Dynamic Link Libraries (DLLs) for Zemax that function as custom surface types (sequential) and custom objects (non-sequential). These DLLs model a grid of flat, rotationally symmetric conic, or cylindrical surfaces.

Key Technical Components:

Parameterization: The DLLs use a parameter set motivated by fabrication processes (specifically the "stack of slices" method used by Bertin Technologies). Parameters include:
- Surface Shape: Curvature ( $c_v$ ) and conic constant ( $\kappa$ ).
- Orientation: Off-axis angles ( $\alpha, \beta$ ) or rotations ( $\gamma$ ) about global axes.
- Position: Translations ( $z_p$ ) and rotation axes offsets ( $s_{xy}, s_{xz}, s_{yx}, s_{yz}$ ).
- Layout: Parameters to define "sections" (groups of slices polished together) and the spacing of slices within and between sections.
Modes of Operation:
- Standard Mode: Exploits the periodicity of most slicers. Users define global parameters (e.g., central angles, spacing increments), and the DLL automatically calculates parameters for thousands of slices.
- Custom Mode: Allows users to manually define parameters for every individual slice via a text file, enabling the modeling of arbitrary, non-periodic designs.
Ray Tracing Algorithm (Sequential):
- Instead of iterating through every slice blindly, the DLL uses a bounding box approach. It calculates global extrema ( $z_{min}, z_{max}$ ) and row offsets to determine the specific range of slices a ray could possibly intersect.
- The algorithm transforms the incoming ray into the local coordinate system of the suspected slice, solves for the intersection analytically (or iteratively for complex cases), and transforms the output ray back to global coordinates.
- It explicitly handles wall collisions (steps between slices) and gaps, which is crucial for stray light and diffraction analysis.
Non-Sequential Implementation:
- The object is approximated by a mesh of triangular facets (similar to STL files) to provide an initial ray-intersection guess.
- Once a ray hits a facet, the DLL provides the exact analytical solution for the transfer distance and surface normal, ensuring high accuracy despite the mesh approximation.
- It allows independent definition of optical properties (coatings, scattering) for slice surfaces, walls, and gaps.

3. Key Contributions

Unified Modeling: The DLLs allow an entire image slicer (with hundreds of slices) to be modeled in a single Zemax configuration, eliminating the need for nested configurations.
Diffraction Analysis: By keeping the system in sequential mode, the DLLs enable Physical Optics Propagation (POP) to analyze the combined diffraction of the entire IFU, a capability previously impossible with multi-configuration workarounds.
Speed Optimization: Ray tracing is significantly faster (seconds vs. minutes/hours) compared to multi-configuration designs, facilitating rapid tolerancing and thermal analysis.
Flexibility: The "Custom" mode supports arbitrary grids, allowing the modeling of free-form surfaces, microlens arrays, and complex pupil re-imaging systems that do not follow standard periodic patterns.
Fabrication-Aware Design: The parameters are mapped directly to manufacturing constraints (e.g., polishing a stack of slices), bridging the gap between optical design and fabrication.

4. Results and Validation

Accuracy: The DLLs were validated by comparing spot diagrams and ray traces against native Zemax surfaces using coordinate breaks. The results were identical.
Case Study (SPECTRE): The DLLs successfully replicated the design of SPECTRE, a facility spectrograph for the NASA Infrared Telescope Facility (IRTF) featuring a 36-slice image slicer.
- The DLL model reduced the file complexity from multiple configurations to a single file.
- It enabled the simulation of diffraction patterns (Airy rings) interacting with adjacent slices, revealing stray light paths that were invisible in the traditional multi-configuration model.
- Ray trace update times dropped from prohibitively slow to 1–2 seconds.
Stray Light Modeling: The non-sequential DLLs successfully modeled scattering and grazing incidence reflections off the "steps" between slices, demonstrating the ability to define unique scattering properties for different surface groups.

5. Significance and Future Work

Impact on Optical Design: This work removes a major bottleneck in the design of integral field spectrographs. It allows optical engineers to optimize image slicers directly within Zemax, including diffraction effects and stray light analysis, which were previously out of reach.
Extensibility: The code is modular and can be extended to model other discontinuous surface grids, such as digital micromirror devices (DMDs) or complex wavefront sensors.
Future Directions:
- Optimization Integration: Future versions aim to pass parameters through "Data Surfaces" to allow Zemax's built-in optimization engine to adjust slicer parameters directly.
- Design Automation: The ultimate goal is to develop a tool that starts from system-level requirements (FOV, F-ratio, sampling) and automatically generates the DLL parameters for a manufacturable IFU.
- CAD Generation: The ability to export human-readable CAD files for vendors is a planned improvement to streamline the manufacturing handoff.

In summary, Ellen Lee's work provides a robust, high-performance software solution that fundamentally changes how image slicer IFUs are designed and analyzed in Zemax, enabling more accurate, efficient, and comprehensive optical system development for next-generation astronomical instruments.