🔬 applied physics

Massive coherent equipartition of light by the geometric phase of null space

This paper presents a scalable, purely geometric scheme that achieves massive coherent equipartition of light from a single on-chip source by leveraging the geometric phases of a null space, experimentally demonstrated with one-to-nine distribution on a glass-based photonic chip.

Original authors: Xiangrui Hou, Dongyi Wang, Fangyu Wang, Congwei Lu, Zhaoju Yang, Guancong Ma

Published 2026-02-05

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Xiangrui Hou, Dongyi Wang, Fangyu Wang, Congwei Lu, Zhaoju Yang, Guancong Ma

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 have a single, powerful flashlight. In the world of high-tech optics, scientists often need to take that single beam of light and split it into many smaller beams that are perfectly synchronized—like a choir of singers all hitting the exact same note at the exact same time. This is crucial for things like quantum computing and advanced sensors.

However, current technology is a bit clumsy at this. If you try to split light using standard tools, the resulting beams often get out of sync, like a choir where everyone is singing slightly off-key or at different speeds. If you try to use many separate flashlights instead, they naturally drift out of step with each other.

The Paper's Big Idea: The "Magic Dance Floor"

This research introduces a clever new way to split one beam of light into many perfectly synchronized beams using a "magic dance floor" made of glass.

Here is the analogy:

The Problem with Standard Splitters: Imagine trying to split a group of dancers into two lines. If you just push them apart, they might stumble, trip, or end up with different rhythms. In physics terms, this is "dynamic phase" noise—it messes up the timing and coordination of the light.
The "Null Space" (The Safe Zone): The researchers discovered a special mathematical "safe zone" (called a null space) where light waves can exist without accumulating any of that messy timing noise. Think of this as a dance floor where the music is perfectly still, so the dancers can move without ever getting out of step.
The Geometric Phase (The Dance Move): Instead of pushing the light, the researchers guide it through a specific, twisting path. Imagine the light is a dancer moving in a circle on this safe floor. Because of the shape of the path (a "geometric phase"), the dancer doesn't just move forward; they rotate their position relative to the other dancers.
- If the dancer spins a specific amount (a specific angle), they can swap places with a partner.
- If they spin a different amount, they can split into two dancers, perfectly synchronized.
- If they spin in a complex pattern, they can split into many dancers, all still holding hands and moving in perfect unison.

What They Actually Did

The team built a tiny chip made of glass with a series of microscopic tunnels (waveguides) carved into it using a laser. They didn't just carve straight lines; they curved the tunnels in a very specific, twisting pattern.

The Experiment: They shot a single beam of red laser light into the middle of this tunnel system.
The Result: As the light traveled through the twists and turns, it didn't just split randomly. It followed the "dance moves" of the geometric phase.
- They successfully split one beam into two perfectly synchronized beams.
- They split one beam into five perfectly synchronized beams.
- They even combined these tricks to split one beam into nine perfectly synchronized beams.
The Scale: They also designed a theoretical blueprint to show how this could be scaled up to split one beam into 41 (or even more) beams, all staying perfectly in sync.

Why This Matters (According to the Paper)

The paper claims this is a "purely geometric" solution. Because the light stays in that special "safe zone" (the null space), it never picks up the "noise" that usually ruins synchronization.

Think of it like this: If you want a group of people to walk in perfect lockstep, you don't tell them to "walk fast" or "walk slow" (which leads to mistakes). Instead, you give them a specific, pre-planned path to walk. As long as they follow the path, they will naturally end up in perfect step, no matter how complex the path is.

This work proves that we can build a "traffic controller" for light on a tiny chip that can take one source and distribute it to many destinations without losing the perfect coordination required for advanced technologies like quantum computing.

Technical Summary: Massive Coherent Equipartition of Light by the Geometric Phase of Null Space

Problem Statement
The development of integrated photonic systems for communications, sensing, and computing relies heavily on coherent light sources. A significant bottleneck exists in generating and distributing coherent light from a single on-chip source across multiple channels with high phase stability. Current integrated lasers typically operate in isolation, leading to random phase factors between different sources that compromise phase-sensitive applications like quantum optics. Conversely, conventional splitting technologies, such as multi-mode interferometers (MMIs) and directional couplers, lack inherent phase stability, failing to guarantee the precise phase coherence required for advanced photonic applications.

Methodology
The authors propose a purely geometric scheme to achieve massive, coherent, one-to-many equipartition of light. The core mechanism leverages the geometric phases associated with the null space (kernel) of a chiral-symmetric Hamiltonian.

Theoretical Framework: The system is modeled by a real-valued, chiral-symmetric Hamiltonian $H(\lambda)$ with a null space spanned by degenerate eigenmodes at zero eigenvalue. Under adiabatic evolution driven by a control parameter $\lambda$ , the zero modes evolve according to a unitary operator $\mathcal{U}$ . Because the system possesses time-reversal symmetry and real eigenvectors, the evolution is restricted to the special orthogonal group $SO(N)$, where $N$ is the dimension of the null space.
Geometric Phase: Crucially, since the modes are at zero energy, the evolution operator generates no dynamic phase; the resulting transformation is purely geometric. The evolution of the null space maps to a real-space rotation described by $SO(N)$. By engineering the system parameters to trace a specific closed loop in parameter space, the authors can control the solid angle ( $\Omega$ ) enclosed, thereby dictating the final superposition state of the light.
Experimental Implementation: The scheme is realized using a waveguide array fabricated on a Corning 7980 glass substrate via femtosecond laser direct-writing. The coupling coefficients between waveguides are modulated along the propagation direction ( $z$ ) by varying the center-to-center spacing. This modulation creates the required adiabatic evolution path on a parameter space sphere.

Key Contributions and Results
The paper demonstrates the scalability and precision of this geometric approach through several experimental realizations:

One-to-Two Equipartition: Using a four-waveguide system (a $1 \times 3$ Hamiltonian model), the authors achieved a 50-50 split of light injected into a single waveguide. By enclosing a solid angle of $\Omega = \pi/4$ , the light was equally distributed to two output waveguides with coherent in-phase or out-of-phase relationships depending on the injection site.
One-to-Five Equipartition: An 11-site model with a three-dimensional null space ($SO(3)$) was utilized to split light from a central site into five output channels (sites 5–9) with equal intensity. Experimental results showed intensity variations of less than 5.6% across the outputs.
One-to-Nine Equipartition: By cascading a one-to-five unit with four one-to-two units, the authors achieved a near-uniform distribution of light across nine output waveguides.
Scalable 1-to-41 Design: The authors propose a modular architecture to extend the splitting count to $N=41$ . This design combines one unit of one-to-five splitting, twelve units of one-to-three splitting, and twelve units of one-to-two splitting. Theoretical calculations confirm that by carefully adjusting the solid angles at each step, the incident light can be evenly distributed across all 41 output sites while maintaining phase coherence.

Significance
This work establishes a versatile and scalable platform for integrated coherent light sources. By exploiting the geometric phases of null spaces, the proposed scheme overcomes the phase instability limitations of conventional splitters and the isolation issues of independent lasers. The purely geometric nature of the control ensures that coherence is not contaminated by dynamic phases. The authors assert that this framework paves the way for integrated photonic applications requiring high-fidelity coherent light distribution, specifically citing potential impacts on quantum photonics and optical computing. The ability to scale from simple splits to massive $N$ -channel distributions on a single chip represents a significant step toward complex, on-chip quantum-optical systems.

More like this