🔬 optics

Geometric scaling limits of phase-only control in multimode coherent systems

This paper establishes that restricting coherent wave control to relative phases confines system dynamics to a compact geometric manifold, revealing universal scaling limits where increased stationary states degrade localization contrast, and demonstrates a practical, calibration-free method to identify robust operating regimes in multimode photonic systems without requiring complex system reconstruction.

Original authors: Harikumar K Chandrasekharan

Published 2026-06-04

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Harikumar K Chandrasekharan

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 focus a flashlight beam to a single, sharp point on a wall. In the ideal world of physics, you could adjust every tiny detail of the light waves—how strong they are and exactly when they peak—to make that point perfectly bright.

But in the real world, especially with complex systems like fiber optic cables or acoustic chambers, you often can't control the strength of the waves. You can only control their timing (their "phase"). You can tell the waves to arrive a little earlier or a little later, but you can't make them louder or softer.

This paper explores what happens when you are forced to work with only these timing adjustments. The author, Harikumar Chandrasekharan, uses geometry and math to show that this limitation creates a very specific, predictable landscape of possibilities.

Here is the breakdown of the paper's findings using simple analogies:

1. The "Hilly Landscape" of Timing

Imagine the different ways you can set the timing of your waves as a giant, rolling landscape.

The Terrain: Because you can only adjust timing, this landscape isn't an open field; it's a closed, finite shape (mathematically, a "torus" or a donut shape).
The Valleys and Peaks: On this landscape, there are specific "peaks" where the waves line up perfectly to create a bright, focused spot. These are called stationary states.
The Basins: Around each peak, there is a "basin" or a bowl. If you are anywhere inside this bowl, the light is still fairly focused. If you step out of the bowl, the light scatters and becomes a messy blur.

The Key Insight: Even though you can't control the strength of the waves, the geometry of this "timing landscape" guarantees that these focused peaks exist. They are like hidden treasure spots on a map.

2. The Size Problem: Why Bigger Isn't Always Better

You might think, "If I add more waves (more modes) to the system, I should be able to focus the light even better." The paper says: Not necessarily.

Here is the catch:

The Leak: In any real system, some of the energy "leaks" out into places you can't control (like stray vibrations or light going into the wrong part of a fiber).
The Scaling Law: As you add more waves to your system, the number of "treasure spots" (peaks) increases. However, the quality of those spots gets worse.
The Analogy: Imagine trying to fill a bucket with a hose. If you have a small bucket (few modes), you can fill it easily. If you have a massive, complex bucket with holes in it (many modes + leakage), adding more water (more modes) doesn't make the bucket fuller; it just makes the water splash out more. The "contrast" (how bright the spot is compared to the background) drops because the energy leaks into the uncontrolled parts of the system.

The paper proves that there is a universal "ceiling" on how well you can focus the light. No matter how clever your algorithm is, if you have leakage and only control timing, adding more complexity eventually hurts your performance.

3. Imperfections Just "Softens" the Hills

Real life is messy. There are imperfections in the equipment, losses in the cables, and slight misalignments.

The Effect: These imperfections don't destroy the treasure spots. Instead, they make the "bowl" around the peak softer and shallower.
The Result: It becomes easier to accidentally step out of the "good" zone. The focused spot is still there, but it's less robust against tiny jitters or noise. The paper shows that you can measure how "steep" the sides of these bowls are to know how stable your system is.

4. The "No-Map" Experiment

Usually, to fix these systems, scientists have to build a complex mathematical map of the entire system (called a "transmission matrix") and run endless computer simulations to find the perfect settings. This takes a long time and requires perfect knowledge of the hardware.

The author demonstrates a much simpler way using a device called a photonic lantern (a tool that combines light from multiple fibers).

The Method: Instead of building a map, they simply turned the knobs (voltage) that control the timing and watched what happened.
The Discovery: By scanning the knobs roughly, they could see the "plateaus" where the light stayed bright and stable. They found the best settings just by looking for these stable "flat spots" in the data.
The Benefit: You don't need to know the inner workings of the machine or run complex optimizations. You just need to scan the controls and find the spots where the light stays steady.

Summary

This paper tells us that when you are limited to controlling only the timing of waves in a complex system:

Good spots exist: There are specific settings where the waves focus perfectly, and these are guaranteed by the geometry of the system.
Bigger is harder: Adding more waves increases the number of good spots, but makes each one weaker because of unavoidable energy leaks.
Imperfections soften the target: Real-world flaws make these good spots less stable, but they don't make them disappear.
Simple works best: You don't need complex math to find these spots. A simple, rough scan of the controls is enough to find the stable, high-performance settings.

This approach offers a practical, "calibration-free" way to operate complex optical and wave systems without needing to understand every single detail of how they work.

Technical Summary: Geometric Scaling Limits of Phase-Only Control in Multimode Coherent Systems

Problem Statement
In multimode coherent wave systems (e.g., optical fibers, acoustic cavities, microwave resonators), controlling energy distribution often relies on phase-only actuation due to practical constraints. While substantial progress has been made using wavefront shaping and transmission-matrix approaches, these typically assume unconstrained control over complex amplitudes or require detailed system knowledge and iterative optimization. A fundamental gap remains in understanding the physical limits of systems where control is restricted strictly to relative phases. Specifically, it is unclear how the geometry of the control space dictates the existence, stability, and scalability of localized interference states, and how imperfections (loss, mixing, leakage) affect these states without full system inversion.

Methodology
The authors establish a minimal geometric framework for phase-only control in coherent multimode systems, treating the control space as a compact manifold rather than a high-dimensional unconstrained space.

Geometric Framework: The system is modeled as a superposition of $N$ spatial modes with fixed amplitudes and tunable relative phases. Removing the global phase redundancy confines the control parameters to an $(N-1)$ -dimensional torus ( $T_{N-1}$ ).
Localization Functionals: The authors define scalar localization functionals (e.g., intensity at a target point) over this compact phase space. The resulting landscape is smooth and compact, enforcing the existence of isolated stationary points (basins of attraction) rather than continuous steerable states.
Stability Analysis: Robustness is quantified by the local curvature (Hessian eigenvalues) of the localization landscape at stationary points. This "basin stiffness" predicts noise resilience.
Imperfection Modeling: Non-idealities (loss, residual mixing, leakage) are modeled as smooth deformations of an effective Gram operator ( $G_{eff}$ ). This approach links physical disorder to geometric deformation of the phase-space landscape.
Scaling Law Derivation: The authors derive a geometric scaling law relating achievable localization contrast ( $\eta$ ) to the number of controlled modes ( $N$ ) and the leakage fraction ( $\ell$ ) into uncontrolled degrees of freedom, defined by the mismatch parameter $\Xi \equiv \ell(N-1)$ .
Experimental Realization: The theory is tested using a telecom-wavelength (1550 nm) three-mode photonic lantern. The experiment employs coarse, voltage-driven phase scans (where voltage proxies relative phase) to map the phase-space landscape without transmission-matrix reconstruction or adaptive optimization.

Key Contributions

Geometric Origin of Limits: The paper demonstrates that restricting control to relative phases confines dynamics to a compact manifold, which generically produces discrete, isolated stationary interference configurations. This contrasts with unconstrained control landscapes.
Universal Scaling Constraint: A fundamental trade-off is identified: while the number of stationary states increases with system dimensionality ( $N$ ), the achievable localization contrast degrades due to leakage into uncontrolled degrees of freedom. The degradation follows the scaling law $\eta_{max} \lesssim 1 - C\Xi$ , where $\Xi = \ell(N-1)$ .
Basin Softening: Imperfections do not eliminate stationary states but "soften" the basin structure (reducing curvature), thereby reducing robustness and contrast. This softening is a geometric consequence of the mismatch between the controlled phase manifold and the uncontrolled leakage subspace.
Calibration-Free Identification: The authors show that stable, high-efficiency operating regimes can be identified experimentally through coarse phase scans. Stable basins appear as plateau regions in energy traces, allowing for the direct mapping of robust operating points without system inversion or iterative optimization.

Results

Simulation Results: Numerical simulations of a four-mode system illustrate the phase-space landscape, showing multiple distinct basins (B1–B4) corresponding to localized output patterns. As imperfection strength increases, basin centers shift, depths split, and eventually, basins merge or lose distinct identity in a leakage-dominated regime.
Scaling Behavior: In photonic lantern simulations, increasing the number of modes ( $N$ ) while maintaining a fixed leakage fraction leads to a systematic degradation of the mean localization index. The number of accessible basins scales linearly with $N$ , but their quality (contrast and stiffness) degrades as $\Xi$ increases.
Experimental Findings: In the three-mode photonic lantern experiment, coarse voltage scans successfully mapped the phase-space structure. Distinct localized states (Spots 1–3) were identified as separate basins with varying efficiencies and spatial footprints. Temporal analysis revealed that "Spot 1" offered the highest efficiency combined with strong temporal stability (plateau behavior), confirming the existence of robust operating regimes accessible without full calibration.

Significance and Claims
The paper claims to establish a practical, calibration-free framework for constrained multimode coherent control. Its primary significance lies in shifting the perspective from algorithmic optimization to geometric constraints:

Fundamental Limits: It clarifies that the inability to achieve perfect localization in large systems under phase-only control is a geometric necessity arising from dimensional mismatch, not merely a failure of optimization algorithms.
Robustness Metric: It introduces basin curvature as a direct, measurable metric for noise robustness, accessible via simple phase perturbations and scalar readouts.
Broad Applicability: The framework is presented as platform-independent, applicable to optical, microwave, acoustic, and finite-dimensional quantum systems where control is restricted to relative phases on a compact manifold.
Practical Utility: The experimental demonstration proves that robust operating regimes can be found directly through measurement, bypassing the need for complex transmission-matrix reconstruction, which is particularly valuable for systems subject to environmental drift and fabrication imperfections.

The authors maintain a modest stance, noting that while stationary configurations persist under imperfections due to compactness, their utility is ultimately limited by the geometric scaling law governing contrast and robustness. They do not claim to overcome these limits but rather to characterize and navigate them.