🔬 optics

Wavelength-Dependent Evolution of Full-Field Transfer Matrices in Photonic Lanterns

This paper provides a quantitative physical explanation for the wavelength-dependent evolution of photonic lantern transfer matrices by demonstrating that differential modal phase accumulation in the multimode section is the dominant mechanism, a finding supported by high-fidelity holographic measurements and a predictive propagation model.

Original authors: Caleb Dobias, Miguel Römer, Swati Bhargava, Tara Crowe, Liza F. Quinn Reyes, David Smith, Matias Barzallo, Daniel Cruz-Delgado, Sergio Leon-Saval, Stephanos Yerolatsitis, Miguel A. Bandres, Stephen S.

Published 2026-04-27

📖 3 min read☕ Coffee break read

CC BY 4.0

Original authors: Caleb Dobias, Miguel Römer, Swati Bhargava, Tara Crowe, Liza F. Quinn Reyes, David Smith, Matias Barzallo, Daniel Cruz-Delgado, Sergio Leon-Saval, Stephanos Yerolatsitis, Miguel A. Bandres, Stephen S. Eikenberry, Rodrigo Amezcua-Correa

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

The "Prism-Pipe" Problem: Making Sense of Light’s Secret Map

Imagine you have a group of 19 different colored flashlights (the Single-Mode Fibers). You want to shine them all into a single, large, thick glass pipe (the Multimode Fiber).

In a perfect world, you’d want to know exactly how each flashlight's beam spreads out inside that big pipe so you could "undo" the mess later. This "map" of how light moves from the small inputs to the big output is what scientists call a Transfer Matrix.

The problem? This map isn't permanent. It changes depending on the color (wavelength) of the light you use. If you use red light, the map looks one way; if you use blue light, the map looks completely different. For engineers, this is like trying to use a GPS that only works if you are driving at exactly 60 mph—the moment you speed up or slow down, the map becomes useless.

The Discovery: The "Dancing Modes" Effect

The researchers at the University of Central Florida wanted to know why this map changes so much. They used a high-tech technique called holography (essentially taking 3D light pictures) to watch how the light behaved across different colors.

They discovered that the "map" doesn't actually change much in terms of where the light goes (the brightness/amplitude), but it changes wildly in terms of timing (the phase).

The Analogy: The Marathon Runners
Think of the different ways light can travel through the big pipe as different runners in a marathon.

Some runners take the straight path in the middle (fast modes).
Some runners zig-zag near the edges (slow modes).

As the race goes on (as the light travels down the length of the pipe), the fast runners pull ahead of the slow runners. Now, here is the kicker: the speed of these runners changes depending on the color of their jerseys.

If you change the color of the light (the wavelength), the runners change their pace. By the time they reach the finish line (the end of the fiber), the "lineup" of runners looks totally different. This "stretching out" of the runners is what causes the Transfer Matrix to evolve and change.

The "Aha!" Moment: Control the Length, Control the Map

The most important part of this paper is that the scientists found a way to predict and control this chaos.

They realized that the "mess" is mostly caused by the length of the big pipe after the light has finished squeezing through the narrow part.

If you want a "Broadband" device (a device that works perfectly for all colors at once, like a high-speed internet cable), you should make the big pipe as short as possible. This prevents the "runners" from spreading out too far, keeping the map stable.
If you want a "Spectroscopic" device (a device that can tell colors apart, like a high-tech sensor), you should make the big pipe longer. This allows the runners to spread out significantly, creating a unique "fingerprint" for every color.

Why does this matter?

This isn't just about playing with light; it's about building better tools for the future. This research helps engineers design:

Better Lidar: Helping self-driving cars "see" more clearly.
Better Telescopes: Allowing astronomers to capture faint light from distant stars more accurately.
Better Medical Imaging: Helping doctors see tiny details inside the human body.

By understanding the "secret map" of light, we can finally stop fighting the chaos and start using it to our advantage.

Technical Summary: Wavelength-Dependent Evolution of Full-Field Transfer Matrices in Photonic Lanterns

Problem Statement

Photonic lanterns (PLs) are waveguide devices that interface an array of single-mode fibers (SMFs) with a multimode fiber (MMF). While "mode-selective" lanterns are designed to maintain a one-to-one mapping between inputs and modes, "non-mode-selective" lanterns—which are easier to scale to high mode counts—exhibit a complex, often seemingly random mapping between SMF ports and MMF guided modes.

Recent experimental studies have noted that the transfer matrix (TM)—the complex-valued mathematical description of this mapping—evolves significantly with wavelength. However, a quantitative physical explanation for why and how this spectral evolution occurs has been lacking. Understanding this evolution is critical for broadband applications (like telecommunications) where stability is needed, or spectroscopic applications where wavelength sensitivity is desired.

Methodology

The researchers employed a combination of advanced experimental measurement and numerical modeling:

Experimental Measurement (Off-Axis Digital Holography):
- The team used a 19-port, non-mode-selective SMF-based PL designed for the 1550 nm band.
- To capture the full complex-valued field (both amplitude and phase), they utilized a Mach-Zehnder off-axis holographic setup.
- They scanned a tunable laser from 1525 nm to 1575 nm at 1 nm intervals, measuring the output field for each of the 19 ports.
- Data Processing: They used the "Fourier Fine-Binning" method to achieve sub-pixel accuracy in centering the holographic twin images, preventing artificial phase jumps. They then decomposed the retrieved electric fields into a basis of Linearly Polarized (LP) modes using a normalized inner product.
Numerical Modeling (Taper Reference Frame Method):
- To isolate the cause of wavelength dependence, they developed a propagation model.
- The model treats the PL as two distinct stages: a taper region (where mode mixing occurs) and a multimode section (where modes propagate with specific phase constants).
- They simulated the TM evolution by applying a diagonal phase matrix $e^{j\beta(\lambda)L}$ to the initial TM, where $\beta$ represents the modal propagation constants and $L$ is the length of the MMF section.

Key Contributions

Direct Full-Field Characterization: Provided high-fidelity, complex-valued measurements of the encoding TM across a 50 nm bandwidth.
Physical Mechanism Identification: Demonstrated that the dominant driver of spectral evolution is differential modal phase accumulation within the multimode section of the lantern.
Predictive Framework: Established a mathematical link between the physical geometry (specifically the length of the MMF section) and the rate of the TM's spectral decorrelation.

Results

Phase vs. Amplitude Evolution: The TMs show rapid evolution in phase but relatively slow variation in amplitude. Correlation analysis showed that while amplitude correlations remain high (~88% over 9 nm), complex-valued correlations (which include phase) drop significantly faster.
Sinc-like Correlation: The correlation between TMs at different wavelengths follows a sinc-like pattern. This pattern is primarily governed by the wavelength difference ( $\Delta\lambda$ ) rather than the absolute wavelength.
Impact of MMF Length: Simulations confirmed that increasing the length ( $L$ ) of the multimode section systematically accelerates the phase evolution, causing the correlation to drop more sharply as wavelength changes.
Taper Effects: While the MMF section dominates phase evolution, the researchers observed wavelength-dependent "mode scrambling" (amplitude decorrelation) within the taper itself, likely due to fabrication imperfections.

Significance

This work moves photonic lantern characterization from empirical observation to predictive engineering. By understanding that the MMF length controls the "speed" of the transfer matrix's evolution, designers can now:

Enhance Broadband Stability: Minimize the MMF section to ensure the TM remains constant over wide bandwidths (ideal for multiplexers).
Enable Wavelength Discrimination: Utilize longer MMF sections to create highly sensitive, wavelength-dependent responses (ideal for spectroscopy or sensing).

This provides a fundamental design framework for tailoring the spectral response of photonic lanterns to meet specific application requirements.