Density-Driven Mott Transition and Exciton Dissociation

Hello! In our previous lesson, we completed our picture of a single exciton by incorporating the electron-hole exchange interaction, which gives rise to its fine structure. Now we are ready to move from the single-particle to the many-particle limit and ask a crucial question: what happens when we create a high density of excitons?

Introduction

Today, we'll explore what happens when excitons are no longer isolated. As their density increases, the interactions between them become dominant, leading to a dramatic change in the state of the semiconductor. This lesson will fulfill the learning outcome: to describe the density-driven Mott transition where excitons dissociate into an electron-hole plasma.

This phenomenon, first predicted by Sir Nevill Mott, is a classic example of a metal-insulator transition driven by electron-electron interactions. We will see how a gas of electrically insulating excitons can abruptly transform into a sea of free electrons and holes that behaves like a metal.

To build a comprehensive understanding, we will approach this transition from three complementary perspectives:

The intuitive physical mechanism: How screening of the Coulomb force by the charge carriers themselves leads to the "unbinding" of excitons.
A thermodynamic viewpoint: We will treat the exciton gas and the electron-hole plasma as two distinct thermodynamic phases and study the equilibrium between them.
A formal many-body approach: We will see how this entire physical system can be elegantly mapped onto the repulsive Hubbard model, the canonical model for studying Mott transitions in condensed matter physics.

The Core Mechanism: Screening and the Mott Criterion

At its heart, the Mott transition is driven by the screening of the Coulomb interaction. A single, isolated exciton is a bound state held together by the attractive Coulomb potential, which in a semiconductor is already screened by the dielectric constant of the material, $V (r) = - e^{2} / (4 π ϵ_{0} ϵ_{r} r)$ .

Now, imagine we photoexcite the semiconductor intensely, creating a high density $n$ of electron-hole pairs. Some will form excitons, but at any finite temperature, some will remain as free electrons and holes. This dense gas of mobile charge carriers fundamentally alters the potential experienced by any single electron-hole pair. The surrounding charges rearrange to screen the attraction, transforming the long-range $1/ r$ potential into a short-range one. This is often described by a Yukawa or screened Coulomb potential:

V_{screened} (r) = - \frac{e ^{2}}{4 π ϵ _{0} ϵ _{r} r} e^{- r / λ_{s}}

where $λ_{s}$ is the screening length. The crucial point is that the screening length depends on the density of free carriers, $n$ . In a degenerate plasma, it's given by the Thomas-Fermi screening length, $λ_{T F}$ , which scales as $λ_{s} \propto n^{- 1/6}$ . As the density $n$ increases, the screening length $λ_{s}$ decreases.

When the screening becomes strong enough, the potential well is too shallow and short-ranged to support a bound state. At this point, excitons can no longer exist and they "dissociate" or "ionize" into a free electron-hole plasma.

Mott Metal-Insulator Transition and Screening — This figure illustrates the effect of screening on a potential well. On the left, with a large screening length (low carrier density), the potential is deep enough to host a bound state (insulating phase with excitons). On the right, with a small screening length (high carrier density), the potential is too shallow, and only free, unbound states exist (metallic phase or electron-hole plasma).

This intuitive picture is captured by the Mott criterion. A bound state ceases to exist when the screening length becomes comparable to the size of the bound state itself, i.e., the exciton Bohr radius, $a_{B}$ . A more precise calculation for the Yukawa potential shows that the last bound state disappears when:

λ_{s} \approx 0.84 a_{B}

Relating the screening length to the density, this leads to the famous Mott criterion for the critical density $n_{M}$ :

n_{M}^{1/3} a_{B} \approx 0.25

This simple relation provides a powerful estimate for the density at which an exciton gas will transition into an electron-hole plasma.

Test your understanding!

For Gallium Arsenide (GaAs), the static dielectric constant is $ϵ_{r} \approx 12.9$ and the exciton reduced mass is $μ \approx 0.063 m_{e}$ , where $m_{e}$ is the free electron mass.

Calculate the 1s exciton Bohr radius $a_{B}$ .
Using the Mott criterion, estimate the critical density $n_{M}$ for the Mott transition in units of cm⁻³.

Show answer

The exciton Bohr radius is given by $a_{B} = \frac{4 π ϵ _{0} ϵ _{r} ℏ ^{2}}{μ e ^{2}} = a_{0} \frac{ϵ _{r} m _{e}}{μ}$ , where $a_{0} \approx 0.529$ Å is the hydrogen Bohr radius.
$a_{B} = (0.529 \times 1 0^{- 10} m) \times \frac{12.9 \times m _{e}}{0.063 m _{e}} \approx 108.2 \times 1 0^{- 10} m \approx 10.8 nm$
Using the Mott criterion, $n_{M} \approx (0.25/ a_{B})^{3}$ .
$n_{M} \approx (\frac{0.25}{10.82 \times 1 0 ^{- 9} m})^{3} = (\frac{0.25}{10.82 \times 1 0 ^{- 7} cm})^{3} \approx (2.31 \times 1 0^{5} cm^{- 1})^{3} \approx 1.23 \times 1 0^{16} cm^{- 3}$
So, for GaAs, the Mott transition is expected to occur at a density of roughly $1 0^{16}$ electron-hole pairs per cubic centimeter.

A Thermodynamic Perspective: The Electron-Hole Liquid

The screening argument gives the "why," but a thermodynamic approach describes the "how." We can view the transition as a competition between two phases: an insulating exciton gas (EG) and a metallic electron-hole liquid (EHL). The EHL is a degenerate, high-density plasma where exchange and correlation energies among the electrons and holes become significant, stabilizing the liquid phase.

To determine which phase is stable at a given density and temperature, we must compare their free energies. Let's first look at the energy of the EHL at zero temperature.

Investigating the Phase Behavior of Indirect Excitons

To understand the stability of the EHL, we need to analyze its total energy. The following paper provides a clear breakdown of the energy contributions for a degenerate electron-hole liquid.

Please read the section 'Quantum liquid' (starts on page 5). Focus on the equation for the total energy, E_tot, and its components (Eqs. 1-4). You don't need to dive into the RPA calculation for the correlation energy E_corr (Eq. 5), but understand what each term physically represents: the kinetic energy of the fermionic carriers, the favorable exchange energy, and the correlation energy.

As the paper shows, the total energy per pair in the EHL has a minimum at a specific equilibrium density. If this minimum energy is lower than the energy of an isolated exciton ( $- E_{e x}$ ), the liquid phase is energetically favorable over the gas phase.

At finite temperature, the situation is more dynamic. There is a chemical equilibrium between bound excitons ( $X$ ) and free carriers ( $e^{-} + h^{+}$ ):

e^{-} + h^{+} ⇌ X

The balance is governed by a law of mass action, which determines the ionization fraction $α$ —the fraction of total pairs that are unbound. A system with $α \approx 0$ is an exciton gas, while $α \approx 1$ is a plasma.

Investigating the Phase Behavior of Indirect Excitons

The transition between the exciton gas and the plasma is not instantaneous. At any finite temperature, there is a chemical equilibrium between bound excitons and free carriers. This paper beautifully describes this equilibrium using a law of mass action.

Now, read the subsequent section starting with 'Next we turn to the effect of thermal fluctuations...' (pages 5-7). Pay close attention to the concept of chemical equilibrium and the introduction of the ionization fraction α (Eqs. 8-9). See how α changes with density in Figure 4. This abrupt change is the Mott transition. Finally, the supplementary information sections 'Solving the law of mass action' and 'Maxwell equal-area constructions' (pages 12-13) explain how the most stable α is chosen by minimizing the free energy and how a Maxwell construction is used to find the phase coexistence region. This is a standard procedure in thermodynamics that should be familiar from your statistical mechanics courses.

To summarize the key points from this thermodynamic analysis:

As density increases from a very low value, pairs start to form excitons, and the ionization fraction $α$ decreases.
At a critical density, exchange and correlation effects in the plasma become so favorable that $α$ abruptly jumps towards 1. This discontinuous change in conductivity signals a first-order Mott transition.
This behavior is analogous to a liquid-gas phase transition. Below a critical temperature, there is a region of phase coexistence, where droplets of the dense EHL form within the sparse exciton gas. The system's phase diagram can be mapped out using a Maxwell construction on the chemical potential, as shown in Figure 5 of the paper.

A Many-Body Perspective: Mapping to the Hubbard Model

Your background in advanced quantum mechanics and statistical physics makes you well-equipped for a more abstract and powerful perspective. The exciton Mott transition is a prime example of a transition driven by strong correlations, and it can be elegantly mapped onto the repulsive Hubbard model, the canonical model for such phenomena.

Exciton Mott transition revisited

A powerful theoretical approach is to map the complex problem of interacting electrons and holes onto a simpler, canonical model. This paper shows how the photoexcited semiconductor system can be mapped onto the repulsive Hubbard model, a cornerstone of strongly correlated electron physics.

Please read Section II, 'THE MODEL' (pages 2-3). Follow the chain of reasoning: start from the general semiconductor Hamiltonian (Eq. 1), simplify it to an on-site interaction model (Eq. 2), and then see how the clever transformation in Eq. (4) maps it to the standard repulsive Hubbard model (Eq. 3). The key result is Eq. (5), which connects the density of electron-hole pairs (n_e-h) to the magnetization (m) of the Hubbard model.

Let's quickly trace the logic of this mapping:

The Hubbard model describes particles hopping on a lattice with an on-site interaction $U$ that penalizes double occupancy.
The mapping is: an empty ground state semiconductor maps to a fully spin-polarized state in the Hubbard model at half-filling.
Creating an electron-hole pair is equivalent to flipping a single spin in the Hubbard model.
Therefore, the density of e-h pairs ( $n_{e - h}$ ) is directly related to the deviation from full spin polarization (magnetization $m$ ). Specifically, $2 n_{e - h} = 1 - m$ .
The insulating exciton gas (EG) corresponds to the Mott insulating phase of the Hubbard model.
The conducting electron-hole liquid (EHL) corresponds to the metallic phase of the Hubbard model.
The exciton binding energy is related to the Hubbard interaction strength $U$ .

With this mapping, the problem of the exciton Mott transition becomes equivalent to studying the metal-insulator transition in the Hubbard model as a function of magnetization, a well-studied problem in many-body theory.

Exciton Mott transition revisited

With the mapping established, we can use the known phase diagram of the Hubbard model to understand the exciton Mott transition. This paper presents the phase diagram as a function of physical parameters like exciton binding energy and e-h pair density.

Please read Section IV, 'EXCITON MOTT TRANSITION...' (pages 7-9) and the conclusion in Section V. Pay special attention to Figure 7, which is the main result. Understand how the Hubbard model axes (U, m) are translated back into the physical parameters of the semiconductor (E_ex, n_e-h). The discussion of the different regions and the nature of the transition (phase separation vs. bistability) is the key takeaway.

This analysis provides a rich theoretical explanation for experimental observations. As the paper concludes, the nature of the transition—whether it is continuous (via phase separation) or discontinuous (an abrupt jump)—can depend on the exciton binding energy. This helps explain why different material systems (e.g., bulk semiconductors vs. quantum wells with larger binding energies) can exhibit different behaviors at the Mott transition.

Conclusion

Today, we have explored the density-driven Mott transition from multiple angles, building a robust understanding of this fundamental many-body phenomenon.

Key Takeaways:

The Mott transition is a phase change from an insulating gas of excitons to a metallic electron-hole plasma, driven by increasing carrier density.
The core physical mechanism is the screening of the Coulomb interaction by free carriers, which destabilizes the bound exciton state above a critical density $n_{M}$ , defined by the Mott criterion.
From a thermodynamic standpoint, it is a first-order phase transition below a critical temperature, analogous to a liquid-gas transition, with a region of phase coexistence between the exciton gas and an electron-hole liquid.
In the language of many-body theory, the system can be mapped onto the repulsive Hubbard model, where the exciton dissociation corresponds to the famous metal-insulator transition.
The exact nature of the transition (continuous or discontinuous) can depend on material parameters like the exciton binding energy, explaining the variety of experimental observations.

Preview of the Next Lesson

We have now established the major concepts in the many-body physics of excitons: the origin of their fine structure via exchange, and their ultimate breakdown into a plasma at high density. In the next module, we will return our focus to the interaction of excitons with light. We will derive the optical absorption coefficient for excitons, discover why excitonic transitions are so strong (high oscillator strength), and analyze the characteristic shape of excitonic absorption spectra. This will connect our theoretical framework directly to the experiments you are familiar with, like photoluminescence and absorption spectroscopy.

Recommended resources