Exciton Recombination: Radiative, Non-Radiative, and Lifetime Impact

Hello! Welcome to your next lesson on exciton dynamics.

In our last session, we constructed the drift-diffusion equation, $\frac{\partial n}{\partial t} = \nabla \cdot (D \nabla n - n μ F) - \frac{n}{τ _{life}}$ , which provides a complete description of how an exciton population evolves in space and time. We ended by noting that the final term, representing the decay of the exciton population, was treated as a simple phenomenological constant, $τ_{life}$ .

Today, we will dissect this lifetime term. Our objective is to describe radiative and non-radiative recombination mechanisms (e.g., Auger processes) and their impact on exciton lifetime. An exciton's life ends when its constituent electron and hole recombine. This can happen in several ways, and these competing decay pathways collectively determine the overall lifetime. The total decay rate, $Γ_{total}$ , is the sum of the rates of all individual processes:

Γ_{total} = \frac{1}{τ _{life}} = Γ_{rad} + Γ_{non-rad}

where $Γ_{rad}$ is the radiative rate and $Γ_{non-rad}$ is the non-radiative rate. Let's explore these channels.

An Overview of Recombination Pathways

Before diving into the details, let's get a high-level view of the primary ways an exciton can recombine.

Exciton Recombination Mechanisms in Band Structures and Device Architectures — This figure illustrates the main recombination pathways for excitons. **(a)** shows ideal radiative recombination. **(b)** shows trap-assisted non-radiative recombination, where defects mediate the process. **(c)** depicts non-radiative Auger recombination, a many-body process involving interactions with other charge carriers (trions, biexcitons, or free carriers). The diagrams on the right show where these loss mechanisms might occur within a device structure.

As the image shows, we can broadly classify recombination into two families:

Radiative Recombination: The exciton annihilates, and its energy is released as a photon. This is the desired process for light-emitting devices.
Non-Radiative Recombination: The exciton's energy is dissipated without light emission, typically as heat (phonons) or by exciting other charge carriers.

The efficiency of a material for light emission is determined by the competition between these pathways. Let's examine the physics of each.

1. Radiative Recombination

In radiative recombination, the electron in the conduction band "falls" back into the empty state (hole) in the valence band, emitting a photon with energy approximately equal to the band gap minus the exciton binding energy. The efficiency of this process depends critically on the band structure of the semiconductor.

Direct Bandgap Materials (e.g., GaAs, many TMDs): The conduction band minimum and valence band maximum occur at the same crystal momentum ( $k = 0$ ). An exciton near the band edge has nearly zero center-of-mass momentum. Since a photon also carries negligible momentum, an electron and hole can recombine directly while conserving momentum. This is a first-order process, making it highly efficient. Radiative lifetimes ( $τ_{rad}$ ) are typically short, on the order of nanoseconds.
Indirect Bandgap Materials (e.g., Si, Ge, h-BN): The conduction band minimum and valence band maximum are at different $k$ -vectors. For an exciton to recombine radiatively, it must simultaneously interact with a photon (to release energy) and a phonon (to conserve crystal momentum). This is a second-order process, which is intrinsically much less probable than a direct transition. Consequently, radiative lifetimes in indirect materials are much longer, ranging from microseconds to milliseconds.

Given your interest in rigorous formalism, it's valuable to see how such a second-order process is modeled. The calculation involves time-dependent second-order perturbation theory, considering both the exciton-photon and exciton-phonon interaction Hamiltonians.

Phonon-Assisted Radiative Lifetimes and Exciton ...

This paper, 'Phonon-Assisted Radiative Lifetimes and Exciton...', develops a first-principles framework to calculate these rates. It provides an excellent example of the theoretical machinery required.

Please read the introduction (the first two paragraphs after the abstract) and the beginning of the theory section up to Equation (9). Don't worry about deriving every term, but focus on: Why a second-order process involving both exciton-photon and exciton-phonon interactions is necessary for indirect gap materials. The structure of the total Hamiltonian (Equation 1) and the interaction terms (Equations 2-4), which formalize the coupling to photons and phonons. The final expression for the transition rate (Equation 7), which includes terms from both interband and intraband processes summed over intermediate states S'. This structure is characteristic of second-order perturbation theory.

The key takeaway from the paper is that the rate for an indirect transition involves a sum over all possible virtual intermediate states, connected by the exciton-phonon and exciton-photon interactions. This complexity explains why the process is so much slower than direct recombination.

2. Non-Radiative Recombination

In many situations, especially in materials that are not perfectly crystalline or at high excitation densities, non-radiative processes dominate, severely shortening the exciton lifetime.

a) Trap-Assisted (Shockley-Read-Hall) Recombination

This is the most common non-radiative pathway in real-world materials. Crystal defects—such as vacancies, interstitials, or impurities—can create localized electronic states within the bandgap. These are known as "traps" or "deep levels."

The process occurs in two steps:

The exciton is captured by the trap state. More accurately, either the electron or the hole is first captured by the defect.
The other carrier is then captured, and the pair recombines at the defect site. The energy is released to the lattice as multiple phonons (heat).

The rate of Shockley-Read-Hall (SRH) recombination, $Γ_{SRH}$ , is proportional to the density of trap sites, $N_{t}$ , and the density of excitons, $n$ . In a simplified view, it contributes a term $A n$ to the total decay rate, where $A \propto N_{t}$ . Therefore, improving material purity and crystalline quality is paramount to increasing the exciton lifetime and light emission efficiency.

b) Auger Recombination

Auger (pronounced oh-zhey) recombination is a three-particle process that becomes the dominant non-radiative channel at high carrier densities. Instead of releasing a photon or phonons, a recombining exciton transfers its energy and momentum to a third particle.

For excitons, the most relevant process is exciton-exciton annihilation: two excitons interact via the Coulomb force. One exciton recombines non-radiatively, while the other is dissociated and promoted to a high-energy, unbound electron-hole state. This high-energy pair then rapidly loses its energy by emitting a cascade of phonons.

The process can be summarized as:

Exciton_{1} + Exciton_{2} \to (electron + hole)_{high-energy} + crystal ground state

Since this requires two excitons, the rate for a single exciton is proportional to the exciton density $n$ . The overall rate of change of the density is therefore proportional to $n^{2}$ . This quadratic dependence is a key signature of Auger recombination in experiments.

The following paper provides a rigorous derivation of the exciton-exciton annihilation rate, which should align well with your preference for theoretical depth.

arXiv:1904.10084v3 [cond-mat.mes-hall] 8 Sep 2019

This paper by Chatterjee and Mabuchi provides a detailed analytical derivation of the exciton-exciton annihilation rate in a 2D material, MoS₂. The fundamental steps are general and offer great insight into the mechanism.

Please read sections II, III, and IV of the paper. Your goal is to understand the main steps of the derivation: Section II (Exciton-Exciton Interaction Energy): See how the interaction between two neutral excitons is modeled as a dipole-dipole interaction (Equation 3). This is the perturbing Hamiltonian, H_{ ext{ex-ex}}. Section III (Interaction Hamiltonian Matrix Elements): Understand the core physical process. To apply Fermi's Golden Rule, we need the matrix element (\langle ext{final} | H_{ ext{ex-ex}} | ext{initial} angle). The initial state is two ground-state excitons, (| u, u angle). The final state is the vacuum and one unbound, high-energy exciton, (|fs, \phi_q angle). This section builds the operator representation needed to calculate this. Section IV (Calculating the Annihilation Rate): Follow how Fermi's Golden Rule (Equation 19) is used with the derived matrix elements to find the annihilation rate. Note how the final rate of density change, d\sigma/dt, becomes proportional to \sigma^2 (Equation 36), establishing the characteristic second-order kinetics of the process.

This derivation beautifully illustrates how a many-body Coulomb interaction leads to a non-radiative decay channel whose importance grows rapidly with exciton density.

Test your understanding!

An experiment measures the time-resolved photoluminescence (TRPL) decay of an exciton population after a short laser pulse. The decay curve is non-exponential. When the initial laser power is increased by a factor of 10 (creating 10 times the initial exciton density), the initial decay becomes much faster.

Which recombination mechanism is most likely responsible for this power-dependent behavior, and why?

Show answer

Auger recombination is the most likely cause.

Here's why:

Radiative recombination ( $Γ_{rad} \propto n$ ) and SRH recombination ( $Γ_{SRH} \propto n$ ) are first-order processes in exciton density $n$ . Their characteristic lifetimes, $τ = 1/Γ$ , do not depend on the density itself.
Auger recombination, specifically exciton-exciton annihilation, is a second-order process. The decay rate of the population density is $d n / d t = - C_{Auger} n^{2}$ . This means the effective lifetime gets shorter as the density $n$ increases.
The experimental observation that a 10x increase in initial density leads to a much faster initial decay directly points to a higher-order process like Auger recombination becoming dominant at higher densities.

The Combined Picture: The ABC Model

We can summarize the total decay rate of an exciton population density $n$ by combining all the mechanisms:

\frac{d n ( t )}{d t} = - Γ_{total} n (t) = - (Γ_{SRH} + Γ_{rad} + Γ_{Auger})

Expressing the rates in terms of their density dependence, this is often written as the "ABC model":

\frac{d n}{d t} = - A n - B n^{2} - C n^{3}

In the context of our discussion:

A: Represents first-order non-radiative SRH recombination and first-order radiative recombination. It dominates at low densities.
B: In the context of exciton-exciton annihilation, this term captures the second-order Auger process ( $B \equiv C_{Auger}$ or $α$ in the notation of paper LINK). Note: In some literature concerning free electron-hole plasmas, the B term is used for bimolecular radiative recombination. Context is key.
C: This term typically describes three-particle Auger recombination in a dense electron-hole plasma, which is less relevant when well-defined excitons exist, but becomes important after the Mott transition.

By fitting experimental decay curves to this equation, researchers can extract the coefficients A, B, and C, thereby quantifying the relative importance of each recombination pathway in a given material.

Conclusion

We have now demystified the $τ_{life}$ term from our transport equation. It is not a single number but the result of a dynamic competition between various decay channels.

Key Takeaways:

Exciton lifetime is limited by parallel radiative and non-radiative recombination processes.
Radiative recombination is efficient in direct-gap materials but requires phonon assistance and is much slower in indirect-gap materials.
Non-radiative recombination includes trap-assisted (SRH) processes, which are linear in density and depend on material defects, and Auger recombination, a many-body effect that dominates at high densities with a characteristic quadratic dependence on density.
The overall population decay can often be described by the ABC model, which accounts for the different density dependencies of each recombination channel.

Preview of the Next Lesson

We've seen that phonons play a crucial role in mediating both radiative recombination (in indirect materials) and non-radiative recombination (dissipating energy from Auger processes). Phonons are also central to how excitons relax within the excitonic state ladder before they recombine. In the next lesson, we will analyze phonon-assisted relaxation processes between exciton states using Fermi's golden rule. This will cover how an exciton created in a high-energy state (e.g., n=2) quickly cascades down to the ground state (n=1) by emitting phonons.

Recommended resources