🔬 optics

Azimuthal super-pupil beam engineering for improved fluorescence depletion microscopy

This paper presents and experimentally validates an azimuthal super-pupil beam engineering technique that utilizes a phase-only spatial light modulator to create a tighter, Bessel-like doughnut-shaped depletion field, achieving a 16% reduction in the central doughnut's peak-to-peak distance compared to conventional beams for improved lateral resolution in fluorescence depletion microscopy.

Original authors: Costanza Agazzi, Nick Toledo-García, Estela Martín-Badosa, Mario Montes-Usategui, David Maluenda, Jordi Tiana-Alsina, Rosario Martínez-Herrero, Artur Carnicer

Published 2026-03-31

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Costanza Agazzi, Nick Toledo-García, Estela Martín-Badosa, Mario Montes-Usategui, David Maluenda, Jordi Tiana-Alsina, Rosario Martínez-Herrero, Artur Carnicer

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 trying to take a photo of two tiny fireflies sitting very close together. In a normal camera, the light from each firefly blurs into a single, fuzzy blob because of a rule called the "diffraction limit." You can't see them as two separate points; they just look like one big smudge.

For decades, scientists have been trying to break this rule to see smaller things, like the tiny machinery inside a cell. One popular method is called STED microscopy. Think of it like this: instead of just shining a light to see the fireflies, you shine a special "eraser" light (a depletion beam) that turns off the glow of any firefly except the one dead center. If you can make that "eraser" light have a perfect, tiny hole in the middle, you can isolate just one firefly and see it clearly.

The Problem with the Old "Eraser"
The standard "eraser" light looks like a donut (a ring of light with a dark hole in the middle). However, the hole in the middle of a standard donut is a bit too wide. It's like trying to pick out a single grain of sand from a beach using a bucket with a hole that's slightly too big—you might accidentally scoop up the sand next to it too.

The New Solution: The "Super-Donut"
This paper describes a new way to make that donut-shaped light beam. The researchers didn't just use a standard donut; they engineered a "Super-Donut."

Here is how they did it, using some creative analogies:

1. The Orchestra Conductor (The Pupil)

Imagine the lens of a microscope is a giant orchestra. The light entering the lens is the music. Usually, the musicians (light waves) play in a simple, uniform pattern.
The researchers acted like a conductor who told the musicians to play a very specific, complex rhythm. They didn't just ask for a simple beat; they asked for a "super-oscillatory" pattern. This is a fancy way of saying they arranged the light waves to interfere with each other in a very precise way.

The Result: Instead of a wide, fuzzy hole in the middle of the donut, they created a hole that is 16% smaller and much sharper. It's like shrinking the bucket's hole so you can pick out that single grain of sand perfectly.

2. The Double-Acting Mirror (The SLM)

To create this complex pattern, they used a device called a Spatial Light Modulator (SLM). Think of this as a high-tech, programmable mirror made of millions of tiny pixels.

The Challenge: Light has two "handedness" properties (polarization), like left-handed and right-handed gloves. To make the perfect donut, you need to control both hands independently.
The Trick: The researchers used a clever setup where the light bounces off the mirror twice (a "double-pass").
- First pass: The mirror tweaks the "left-handed" light.
- Second pass: The mirror twists the light and tweaks the "right-handed" light.
- Outcome: By the time the light leaves, it's a perfectly engineered "vector beam" with the exact shape and polarization needed to make that super-tight hole.

3. The Pixel Puzzle (Pixel Pooling)

There was a catch. The mirror is made of tiny pixels, and sometimes neighboring pixels "talk" to each other (a problem called "cross-talk"), which blurs the image.

The Fix: The researchers decided to group the tiny pixels together into bigger "super-pixels" (like gluing four puzzle pieces together to make one big piece).
The Sweet Spot: They found that grouping them in a 6x6 block was the magic number. It was big enough to stop the pixels from interfering with each other, but small enough to still draw the detailed picture they needed.

4. The Test: Tiny Beads

To prove it worked, they didn't just look at the light; they looked at how it affected tiny glowing beads (smaller than a virus).

The Result: When they scanned these beads with their new "Super-Donut" beam, the glowing spot was significantly tighter than with the old standard beam.
The Trade-off: Just like a super-tight donut might have a few crumbs (sidelobes) around the edge, this new beam has some extra light rings around the main hole. However, for this specific job (erasing fluorescence), those extra crumbs don't matter much. The most important thing is that the hole in the center is tiny, which allows for much sharper images.

Why Does This Matter?

This isn't just about making a prettier donut. By making that central hole smaller, scientists can:

See smaller details: They can distinguish between two molecules that are incredibly close together.
Use less power: Because the hole is tighter, they don't need to blast the sample with as much intense light to get the same result, which is better for delicate living cells.
Be flexible: Because they use a programmable mirror, they can change the shape of the "donut" on the fly to suit different experiments.

In a nutshell: The researchers built a smarter, more precise "eraser" for light. By carefully choreographing how light waves interact and using a double-bounce mirror trick, they shrank the hole in their laser donut, allowing microscopes to see the microscopic world with unprecedented clarity.

1. Problem Statement

Fluorescence depletion microscopy techniques, such as Stimulated Emission Depletion (STED) and Reversible Saturable Optical Fluorescence Transitions (RESOLFT), rely on a depletion beam with a well-defined, spatially confined central intensity minimum (a "doughnut" shape) to achieve sub-diffraction resolution.

Limitations of Current Methods: Conventional STED beams often use circularly polarized light with a vortex phase mask. While effective, these can be further optimized. Vector beams (radially or azimuthally polarized) offer tighter focusing at high numerical apertures (NA), but standard azimuthally polarized beams still have a central minimum width limited by diffraction and topological charge.
The Challenge: Achieving a narrower central zero without increasing the depletion power requires engineering the beam's spatial profile. However, methods like super-oscillatory pupil filtering (super-pupils) typically produce strong sidelobes. In imaging, these sidelobes degrade the point-spread function (PSF) and image reconstruction. In depletion microscopy, while sidelobes do not directly blur the image, they redistribute optical energy away from the central ring, potentially reducing the efficiency of fluorescence suppression. The challenge is to engineer a beam that maximizes central confinement while managing the trade-off with sidelobe energy and experimental implementation constraints (e.g., SLM pixel cross-talk).

2. Methodology

The authors propose a design and experimental implementation of an azimuthally polarized, doughnut-shaped depletion beam using super-pupil engineering.

Theoretical Design:
- Framework: The design utilizes the Richards-Wolf vector diffraction theory for high-NA focusing.
- Target Function: The goal is to approximate a focal intensity profile proportional to $|J_1(snr)|^2$ (a Bessel-type function), which offers a tighter central minimum than standard Gaussian or top-hat profiles.
- Optimization: The radial amplitude distribution at the entrance pupil, $\rho g(\rho)$ , is modeled as a truncated expansion of radial Zernike polynomials ( $R_n^1$ ). The coefficients ( $\alpha_n$ ) are optimized using a Differential Evolution (DE) algorithm to minimize the distance between the calculated focal intensity and the target Bessel function.
- Parameters: The optimization was performed over a design window of $L = 12\lambda$ with a maximum Zernike order of $n=7$ and a scaling factor $s=6$ . This combination was selected to balance central confinement, sidelobe suppression, and axial beam length.
Experimental Implementation:
- Hardware: A phase-only Spatial Light Modulator (SLM) in a double-pass configuration was used. This setup allows for the independent modulation of orthogonal polarization components (essential for generating vector beams) using a single device.
- Encoding: To encode complex fields (amplitude and phase) on a phase-only SLM, the Arrizón double-pixel holographic encoding method was employed.
- Optimization of Encoding: To mitigate pixel cross-talk inherent in reflective SLMs (which distorts fine spatial features and degrades the central zero), the authors implemented pixel pooling. They tested pooling factors from $2\times2$ to $8\times8$ and determined that a $6\times6$ configuration provided the optimal balance between suppressing cross-talk and maintaining sufficient spatial resolution to reproduce the designed pupil function.
Characterization:
- Pupil Plane: Verified the intensity distribution and polarization state (Stokes parameters) at the entrance pupil.
- Focal Plane: Used sub-diffraction fluorescent beads (170 nm and 23 nm) to raster-scan the focal plane and reconstruct the Point Spread Function (PSF).
- Noise Reduction: Applied bilateral filtering to 23 nm bead scan data to reduce high-frequency noise and reveal the true depth of the central zero.

3. Key Contributions

Super-Pupil Vector Beam Design: Successfully designed an azimuthally polarized beam where the pupil amplitude is tailored to generate a Bessel-like focal profile, resulting in a significantly tighter central intensity minimum compared to conventional azimuthally polarized beams.
Robust Experimental Platform: Demonstrated a reconfigurable, double-pass SLM system capable of generating complex vector fields. The study provides a practical protocol for mitigating SLM limitations (pixel cross-talk) through controlled pixel pooling ( $6\times6$ ), ensuring high-fidelity reconstruction of the super-oscillatory features.
Quantitative Validation: Provided experimental evidence that the engineered beam reduces the peak-to-peak distance of the central doughnut by ~16% compared to a nominal reference beam, even in the presence of sidelobes.
Trade-off Analysis: Clarified that for depletion microscopy, the presence of sidelobes (which redistribute energy) is an acceptable trade-off if the central confinement is sufficiently improved, as lateral resolution is primarily dictated by the steepness and width of the central zero.

4. Results

Pupil Verification: Measured intensity and Stokes parameter maps at the entrance pupil confirmed the accurate reproduction of the designed radial modulation and azimuthal polarization (linear polarization components $S_1/S_0$ and $S_2/S_0$ reached ~0.9).
Focal Confinement:
- Scanning with 170 nm beads showed the engineered beam's central doughnut peak-to-peak distance reduced from 0.55 $\lambda$ (reference) to 0.46 $\lambda$ (engineered), a 16% improvement.
- Scanning with 23 nm beads (smaller than the central zero) revealed an even sharper on-axis minimum with steeper intensity walls, confirming that larger beads partially filled the zero and underestimated the suppression.
- Post-processing (bilateral filtering) of the 23 nm data brought the experimental profile into close agreement with the theoretical simulation.
Sidelobes: The engineered beam exhibited pronounced sidelobes (characteristic of Bessel-type beams), but these did not prevent the formation of a deep central zero. The study notes that while sidelobes reduce total power available for depletion, the geometric narrowing of the zero offers a path to higher resolution at comparable powers.

5. Significance

This work establishes a viable pathway for adaptive super-resolution microscopy. By combining super-pupil engineering with vector beam control on a programmable SLM platform, researchers can:

Enhance Resolution: Achieve improved lateral resolution without necessarily increasing the depletion laser power, which is crucial for reducing phototoxicity in biological samples.
Flexibility: The reconfigurable nature of the SLM allows for dynamic adjustment of beam profiles (e.g., changing the scaling factor $s$ or Zernike order) to adapt to different samples or imaging depths.
Overcome Hardware Limits: The demonstrated pixel pooling strategy offers a practical solution for generating high-fidelity super-oscillatory fields on standard phase-only SLMs, making advanced beam engineering accessible for experimental setups.

The authors conclude that while future work should address throughput efficiency (e.g., via phase-retrieval algorithms) and 3D axial characteristics, this approach provides a robust, experimentally accessible route to engineering superior depletion fields for next-generation STED and RESOLFT microscopes.