🔬 applied physics

A transferable full-band Monte Carlo framework for complex alloy avalanche photodiodes

This paper presents a transferable, physics-based multiscale full-band Monte Carlo framework that links atomistic tight-binding band structures and alloy-disorder scattering to accurately model avalanche gain and excess noise in complex quaternary alloy avalanche photodiodes, such as Al $_{0.7}$ InAsSb, with minimal calibration.

Original authors: Shafat Shahnewaz, Hannaneh Karimi, Joe C. Campbell, Avik W. Ghosh

Published 2026-10-01

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Shafat Shahnewaz, Hannaneh Karimi, Joe C. Campbell, Avik W. Ghosh

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

Light is a messenger, but sometimes the message is too faint to hear. In the world of fiber optics and deep-space communication, signals can weaken until they are lost in the static of the universe. To catch these whispers, engineers use a special kind of light detector called an avalanche photodiode. These devices act like a chain reaction of falling dominoes: a single particle of light hits the material, knocks loose an electron, and that electron, accelerated by an electric field, smashes into other atoms to release more electrons. This cascade creates a measurable electrical surge from a tiny spark of light. However, this multiplication process is inherently chaotic. The randomness of the collisions introduces a static known as excess noise, which can distort the very signal the device is trying to amplify. For decades, scientists have sought materials that can multiply light signals without adding too much of this unwanted noise, but predicting how a new material will behave has been a guessing game, often requiring expensive trial and error in the lab.

The challenge lies in the complexity of the materials themselves. Modern detectors often use alloys, which are mixtures of different elements, such as aluminum, indium, arsenic, and antimony. When these elements are mixed, they do not always form a perfectly smooth, uniform crystal. Instead, the atoms can arrange themselves in random patterns or in precise, alternating layers, creating a landscape that changes the path of electrons in subtle ways. Traditional models often treat these materials as if they were simple, uniform blocks, ignoring the jagged edges and hidden valleys created by the atomic arrangement. This simplification fails when the material is complex, leading to inaccurate predictions about how much noise the device will generate. To solve this, researchers at the University of Virginia have developed a new way to simulate these devices from the ground up, starting with the position of every single atom.

The team created a comprehensive computer framework that links the atomic structure of a material directly to the behavior of the electrons moving through it. They focused on a specific mixture, aluminum indium arsenide antimonide, which has shown promise for low-noise performance. First, they mapped out the exact arrangement of atoms in the material, accounting for how the local environment around each atom shifts the energy levels available to electrons. This step revealed a detailed map of the energy landscape, showing not just the main path electrons take, but also the side valleys and hidden gaps that appear because of the mixed atomic structure. By understanding this full map, the researchers could see exactly where electrons would get stuck, where they would speed up, and how they would scatter off the irregularities in the crystal.

With this detailed map in hand, the researchers built a simulation that tracks millions of electrons as they race through the material under high voltage. The simulation accounts for every major force acting on the electrons: how they bounce off the random arrangement of atoms, how they interact with heat vibrations in the crystal, and how they gain enough energy to create new electron-hole pairs. Crucially, the model treats the disorder caused by the mixed elements not as a vague background noise, but as a specific, calculable force derived from the differences in the size and charge of the atoms involved. This allows the simulation to predict how the material will behave based solely on its chemical recipe and atomic layout, without needing to guess at hidden parameters.

When the researchers tested their framework against a real-world device, a one-micrometer-thick detector made from this alloy, the results were striking. The computer simulation reproduced the actual performance of the physical device with high accuracy. It correctly predicted how the gain, or the amount of signal amplification, would increase as the voltage rose. More importantly, it matched the measured levels of excess noise, confirming that the model had captured the essential physics of the chaotic multiplication process. The simulation showed that the material's unique atomic structure plays a central role in keeping the noise low, a detail that simpler models would have missed.

The true power of this work is not just in its success with one specific material, but in its ability to be transferred to others. Because the framework builds its inputs directly from the atomic composition and arrangement, it can be applied to any new mixture of these elements without needing to be rebuilt from scratch. If a scientist wants to design a detector using a different combination of elements, they can simply feed the new atomic recipe into the system, and it will generate the necessary physical rules to simulate the device. This provides a clear, documented path for designing future detectors, moving the field away from guesswork and toward a precise engineering of light and matter. The researchers have demonstrated that by respecting the complexity of the atomic world, we can build better tools to hear the faintest signals from the edge of the known universe.

Technical Summary: A Transferable Full-Band Monte Carlo Framework for Complex Alloy Avalanche Photodiodes

Problem Statement
Avalanche photodiodes (APDs) rely on impact ionization to provide internal gain, but this stochastic process introduces excess noise. While the local-field model (McIntyre) and nonlocal dead-space models describe noise behavior, they often fail to capture the complex physics of modern Sb-based alloys like AlInAsSb. These materials exhibit low ionization-coefficient ratios and low excess noise, yet their electronic structure is complicated by alloy composition, growth sequences (random vs. digital alloys), and phenomena such as band folding, mini-gaps, and non-parabolicity. Existing modeling approaches often rely on simplified effective mass approximations or empirical parameter sets that do not transfer easily between different alloy compositions or growth configurations. The challenge lies in developing a unified framework that links atomistic material structure to device-level avalanche statistics without requiring unique, ad-hoc parameterization for every new material system.

Methodology
The authors present a physics-based, multiscale full-band Monte Carlo framework consisting of four sequential stages:

Atomistic Structure Definition: The APD multiplication region is modeled as a specific alloy configuration (e.g., $Al_{0.7}InAsSb$ ), which can be a random alloy or a digital alloy (short-period superlattice). The study specifically considers a digital alloy structure with alternating monolayers of InAs, AlSb, and AlAs on a GaSb substrate.
Environment-Dependent Tight-Binding (EDTB) Calculation: The electronic structure is computed using an environment-dependent $sp^3d^5s^*$ tight-binding model. Unlike bulk effective mass models, this approach determines Hamiltonian parameters based on the local atomic environment (nearest neighbors, bond lengths, and strain). This captures critical features such as $\Gamma$ , $X$ , and $L$ valley ordering, valence-band splitting due to spin-orbit coupling, and mini-gaps induced by the digital alloy potential. The output includes full-band dispersion relations, group velocities, effective mass tensors, and valley offsets.
Composition-Derived Scattering Models: Scattering rates are parameterized directly from the alloy composition and atomic properties, avoiding empirical fitting for scattering mechanisms.
- Alloy-Disorder Scattering: Modeled using a screened atomic-potential approach where the disorder potential is derived from atomic valence differences, covalent radii, and Thomas–Fermi screening. The potential is decomposed into binary contributions weighted by composition.
- Phonon Scattering: Includes polar optical, acoustic, and intervalley phonon scattering. Parameters are interpolated from binary constituents or derived via Vegard's law.
- Impact Ionization: Modeled using a Keldysh rate formula. While the transport inputs (band structure, scattering rates) are derived from first principles, the impact-ionization softness parameters are calibrated to experimental data.
Monte Carlo Transport Simulation: A stochastic solver tracks carrier trajectories through the device under an applied electric field. It simulates free flights, scattering events (elastic and inelastic), and impact ionization events until carriers exit the multiplication region.

Key Contributions

Unified Workflow: The paper establishes a documented, reproducible workflow that connects atomistic configurations to device-level gain and noise statistics.
Transferability: The framework is designed to be transferable to arbitrary zinc-blende ternary and quaternary alloys. Because every material-dependent input (band structure and scattering rates) is generated from atomic composition and configuration via a single defined procedure, the model does not require structural modification for new materials.
Full-Band Resolution: By retaining the full 3D band structure (including mini-gaps and non-parabolicity) rather than using simplified effective masses, the model accurately represents carrier transport in complex digital alloys where folded bands and avoided crossings are significant.
Physics-Driven Scattering: The alloy-disorder scattering model is derived from fundamental atomic properties (valence, radius, screening) rather than treated as a fitting parameter, providing a more physical basis for modeling disorder in quaternary systems.

Results
The framework was applied to a 1 $\mu$ m thick $Al_{0.7}InAsSb$ p-i-n APD.

Gain Characteristics: The simulation reproduced the measured gain versus bias voltage, capturing both the gradual increase at lower biases and the rapid rise near breakdown.
Excess Noise: The calculated excess-noise factor ( $F$ ) matched experimental measurements and remained below the predictions of standard McIntyre local-field models for the observed gain range, consistent with the known low-noise performance of Sb-based APDs.
Calibration: The agreement was achieved with only the impact-ionization softness parameters calibrated; all band structure and scattering inputs were generated directly from the material composition and configuration.

Significance and Claims
The authors claim that the primary value of this work is the establishment of a systematic pipeline for analyzing and designing complex alloy APDs. The framework demonstrates that gain and ionization statistics can be determined by the joint influence of band structure, disorder, and scattering physics without relying on separate empirical parameter sets for each new alloy.

The paper explicitly notes its limitations to maintain modesty regarding its claims:

Validation is currently limited to one specific $Al_{0.7}InAsSb$ multiplication structure.
The electric field is treated as uniform rather than solved self-consistently with Poisson's equation.
The impact-ionization model still relies on calibrated parameters.
The alloy-scattering calculation assumes a random-alloy limit ( $S=1$ ) even when the electronic structure is calculated for a digital-alloy superlattice.

The authors position this work as a foundational step, suggesting that future tests should include comparisons with measured ionization coefficients, different multiplication layer thicknesses, and other alloy systems to further validate the framework's transferability.

Technical Summary: A Transferable Full-Band Monte Carlo Framework for Complex Alloy Avalanche Photodiodes

More like this