🔬 optics

Noise factor of Brillouin amplifiers

This paper presents a comprehensive Hamiltonian-based spatio-temporal coupled mode analysis demonstrating that the noise factor of Brillouin amplifiers can significantly deviate from the traditional approximation of $F \approx 1+n_{th}$ when accounting for phonon propagation, optical loss, and varying scattering geometries in modern systems.

Original authors: John H. Dallyn, Nils T. Otterstrom, Matt Eichenfield, Peter T. Rakich, Ryan O. Behunin

Published 2026-04-15

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: John H. Dallyn, Nils T. Otterstrom, Matt Eichenfield, Peter T. Rakich, Ryan O. Behunin

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 send a whisper across a long, windy hallway. You want to make that whisper louder so the person at the other end can hear it clearly. This is essentially what a Brillouin amplifier does for light signals in fiber optics and computer chips. It takes a faint beam of light (the whisper) and boosts it using sound waves (the wind).

However, there's a catch. The "wind" in this hallway isn't perfectly still; it's constantly jiggling because of heat. These tiny, random jiggles (called thermal fluctuations) get mixed into your whisper, turning it into static. The Noise Factor is simply a score that tells us how much static the amplifier adds to your signal. A lower score is better.

For a long time, scientists had a simple rule of thumb for this score: "The noise is just the heat you started with." They thought that if you amplified the signal enough, the extra noise added by the machine would be negligible, and the only noise left would be the natural heat noise already present. They called this the "Non-Propagating Phonon" (NPP) approximation. Think of it like assuming the wind in the hallway is so heavy and sluggish that it doesn't actually move; it just sits there vibrating in place.

The Big Discovery
This paper says: "That old rule is wrong for many modern devices."

The authors realized that in many new, high-tech systems, the "wind" (the sound waves, or phonons) is actually moving. It travels down the hallway just like the light does. When you account for this movement, the math changes dramatically.

Here is the breakdown using simple analogies:

1. The Old Way (The "Still Wind" Model)

Imagine the sound waves are like heavy, stationary boulders in the hallway. When you shout (the light signal), the boulders vibrate in place and help amplify your voice. Because they don't move, the random jiggles (heat noise) happen everywhere at once, but they don't travel with your voice.

The Result: The old math said the noise factor is roughly 1 + (Heat). It's a predictable, simple formula.

2. The New Way (The "Moving Wind" Model)

Now, imagine the sound waves are like a river flowing down the hallway.

The Signal: Your shout (the light) rides this river downstream. It gets a massive boost because it's surfing the current.
The Noise: The random jiggles (heat) are like pebbles being thrown into the river at random spots.
The Twist: Because the river is flowing, the pebbles thrown at the start of the river get swept along and amplified for the whole trip. But the pebbles thrown at the end of the river only get a tiny boost before the river ends.
The Surprise: In this moving scenario, the "noise" doesn't get amplified as much as the "signal" does. The signal gets a huge head start, while the noise is scattered and diluted.

The Conclusion:
Because the signal and the noise behave differently when the sound waves are moving, the amplifier is actually quieter (has a better noise factor) than the old models predicted.

Why Does This Matter?

For years, engineers designing these amplifiers used the "Old Way" math. They thought, "Oh, the noise will be high, so we need to design around that limit."

This paper tells them: "You can do better!"

Forward vs. Backward: The paper shows that if the sound and light move in the same direction (Forward), the noise reduction is huge. If they move in opposite directions (Backward), it's a bit more complex, but still different from the old rules.
New Devices: Modern chips and fibers are getting so good at controlling sound that the "moving river" effect is now the norm, not the exception.

The Takeaway

Think of this like upgrading from a bicycle to a sports car.

Old Theory: "The speed limit is 50 mph because of the engine's heat."
New Theory: "Actually, if you aerodynamicize the car (account for the moving sound waves), you can hit 100 mph with less wind resistance than we thought."

The authors have provided a new, more accurate map for engineers. If you are building the next generation of ultra-fast internet or super-sensitive sensors, you can't use the old, simplified map anymore. You need to account for the fact that the "wind" is moving, because that movement actually helps you get a clearer, stronger signal with less static.

1. Problem Statement

Stimulated Brillouin Scattering (SBS) is a powerful mechanism for optical amplification and signal processing, offering high tunability and exponential gain. However, the performance of SBS amplifiers is fundamentally limited by noise arising from thermal-mechanical fluctuations (phonons) that imprint onto the optical signal.

Historically, the noise factor ( $F$ ) of Brillouin amplifiers has been approximated as $F \approx 1 + n_{th}$ , where $n_{th}$ is the thermal occupation number of the phonons. This approximation relies on the Non-Propagating Phonon (NPP) assumption, which posits that phonon decay is so rapid that phonon propagation can be neglected (effectively setting the phonon group velocity, $v_{g,b}$ , to zero).

The Gap: As Brillouin photonics advances into chip-scale systems and new material platforms, many emerging devices exhibit significant phonon propagation (long phonon mean free paths). In these regimes, the NPP approximation fails, and the standard $F \approx 1 + n_{th}$ model no longer accurately predicts the noise performance. There is a lack of a general theoretical framework that accounts for phonon propagation, forward/backward scattering geometries, and optical loss simultaneously.

2. Methodology

The authors developed a comprehensive traveling-wave-based Hamiltonian model to describe the noise, coupling, and field evolution in Brillouin amplifiers.

Theoretical Framework:
- The system is modeled using a Hamiltonian $H = H_0 + H_{int}$ , describing the interaction between pump ( $A_p$ ), Stokes ( $A_s$ ), and phonon ( $B$ ) fields.
- The dynamics are governed by Heisenberg-Langevin equations of motion, which include decay rates ( $\Gamma, \gamma_p, \gamma_s$ ) and Langevin noise forces ( $\eta, \xi_p, \xi_s$ ) representing thermal fluctuations and vacuum noise.
- The authors assume an undepleted pump and focus on the noise contribution from the thermal phonon field, which is the dominant noise source in traveling-wave systems.
Analytical Derivation:
- NPP Limit: They first re-derived the noise factor under the NPP approximation ( $v_{g,b} = 0$ ) to validate their model against existing literature ( $F \approx 1 + n_{th}$ ).
- Full Dynamics: They extended the derivation to include non-zero phonon group velocity ( $v_{g,b} \neq 0$ $v_{g, b} \neq = 0$ ).
  - Forward Intermodal SBS: Solved using Fourier and Laplace transforms to handle initial conditions at $z=0$ .
  - Backward SBS: Solved using a Green's function approach due to the counter-propagating nature of the waves and different boundary conditions (Stokes input at $z=0$ , Phonon input at $z=L$ ).
- Noise Factor Calculation: The noise factor $F$ is defined as the ratio of input Signal-to-Noise Ratio (SNR) to output SNR. The output PSD is calculated by integrating the amplified signal and the noise contributions from thermal phonons and optical vacuum fluctuations.

3. Key Contributions

Generalized Noise Theory: The paper provides the first general theoretical treatment of the SBS noise factor that explicitly accounts for phonon propagation. It bridges the gap between traditional fiber-based models and emerging chip-scale systems where phonon lifetimes are long.
Derivation of Exact Noise Factors: The authors derived closed-form analytical expressions for the noise factor in both Forward Intermodal SBS (Eq. 22) and Backward SBS (Eq. 23) regimes, valid for arbitrary values of phonon decay, Brillouin gain, and optical loss.
Identification of Approximation Limits: The study rigorously defines the parameter space (specifically the dimensionless ratio $\Gamma L / v_{g,b}$ ) where the traditional NPP approximation breaks down.
Resolution of Pump Depletion Artifacts: The authors identified that divergences in backward SBS signal amplitudes at low phonon decay are artifacts of the undepleted pump approximation. They demonstrated via Appendix C that including pump depletion restores physical solutions, ensuring the noise factor remains well-behaved.

4. Key Results

The authors analyzed the noise factor ( $F$ ) as a function of three dimensionless variables:

$\Gamma L / v_{g,b}$ : Phonon decay relative to propagation length.
$G$ : Single-pass Brillouin gain.
$\alpha L$ : Optical loss.

Main Findings:

Deviation from $1 + n_{th}$ : In regimes where phonon propagation is significant (low $\Gamma L / v_{g,b}$ ), the noise factor deviates radically from the standard $F \approx 1 + n_{th}$ .
Noise Reduction: Contrary to the assumption that phonon propagation increases noise, the full dynamical models show that phonon propagation can significantly reduce the noise factor compared to the NPP prediction.
- Mechanism: In the NPP limit, thermal noise is generated uniformly along the amplifier length and experiences the full gain. In the propagation regime, the injected coherent signal grows exponentially across the entire length, while thermal phonons (generated locally) only experience partial gain over their specific propagation path. This results in a higher output SNR.
Forward vs. Backward Scattering:
- Forward SBS: Shows a notable decrease in noise factor at high gains when phonon propagation is included.
- Backward SBS: Exhibits complex behavior with "pseudo-gain resonances" due to boundary conditions. While the signal amplitude can diverge in the undepleted pump model, the noise factor remains stable and well-behaved.
Impact of Optical Loss: While optical loss ( $\alpha L$ ) degrades the noise factor, the primary deviation from the NPP model is driven by the phonon propagation dynamics, not just loss.

5. Significance

Design Guidance for Emerging Devices: As researchers develop SBS devices in new materials (e.g., silicon, chalcogenides) and geometries (e.g., acoustoelectric coupling) where phonon lifetimes are long, the traditional $F \approx 1 + n_{th}$ rule is insufficient. This paper provides the necessary tools to accurately predict and optimize the noise performance of these next-generation amplifiers.
Performance Optimization: The finding that phonon propagation can lower the noise factor suggests that engineering systems to maximize phonon mean free paths could lead to ultra-low-noise Brillouin amplifiers, surpassing the limits previously thought to be fundamental.
Theoretical Foundation: The work establishes a robust Hamiltonian framework for analyzing optomechanical noise in traveling-wave systems, applicable beyond just Brillouin scattering to other photon-phonon interaction platforms.

In summary, the paper demonstrates that the "ultimate limit" of Brillouin amplifier noise is not fixed at $1 + n_{th}$ but is a dynamic quantity heavily dependent on the interplay between phonon propagation, gain, and scattering geometry. Ignoring phonon propagation leads to a severe overestimation of noise in modern, high-performance SBS systems.