Introduction
InGaN-based light-emitting diodes (LEDs) are the foundation of modern solid-state lighting. Despite their success, these devices present a physical paradox: they maintain high radiative efficiency even when grown on substrates with high dislocation densities. These structural "cracks" typically act as non-radiative recombination centers that quench light emission. The prevailing explanation for this resilience is carrier localization—the trapping of electrons and holes in local potential minima, which prevents them from reaching defects.
The Random Alloy Model
For years, the scientific community debated whether this localization was caused by "indium clusters"—regions of high indium concentration—or simple statistical fluctuations. By utilizing Atom Probe Tomography (APT) data, researchers have demonstrated that the indium distribution in these quantum wells (QWs) is statistically random, rather than clustered.
As shown in Figure 2, the experimental frequency of indium content closely follows a binomial distribution. This finding allows for a more realistic theoretical treatment using numerical solutions of the effective mass Schrödinger equation, moving away from idealized "cluster" models to a more rigorous statistical approach.
The Potential Energy Landscape
The study models the energy landscape by accounting for four primary physical factors:
- Local band gap fluctuations arising from varying indium content.
- Deformation potentials caused by local strain variations.
- Spontaneous and piezoelectric fields inherent to the wurtzite crystal structure.
- Well width fluctuations (WWFs), representing monolayer variations in the QW thickness.
Carrier Asymmetry and Localization
A key insight from the numerical simulations is the profound asymmetry between electron and hole localization. Holes, characterized by a higher effective mass, are extremely sensitive to local indium fluctuations. They tend to localize in regions approximately 1 nm in diameter where the indium concentration is slightly above the mean.
In contrast, electrons have a much lower effective mass and are less affected by small-scale alloy fluctuations. As illustrated in Figure 4, the electron wavefunction is significantly more delocalized, typically spreading over 7 to 10 nm. Because of the strong internal electric fields in polar GaN, holes are pushed toward one interface while electrons are pushed toward the other. This spatial separation makes electrons particularly susceptible to well width fluctuations (WWFs) at the rougher interface, whereas holes remain largely unaffected by these interface variations.
Results and Validation
The model predicts that WWFs can reduce electron localization lengths by up to one-third in high-indium samples. To validate these findings, the researchers simulated photoluminescence (PL) spectra. For a 25% indium QW, the calculated Full Width at Half Maximum (FWHM) was 69 meV, showing excellent agreement with the experimental value of 63 meV. This suggests the model captures the dominant physics of the system.
Conclusion
This research confirms that the natural randomness of an InGaN alloy is sufficient to explain the carrier localization necessary for efficient LED operation. The statistical "puddles" created by random fluctuations are enough to trap holes, while well width variations provide the necessary localization for electrons. This theoretical framework provides a clear explanation for why InGaN LEDs remain efficient despite significant structural defects.
Summary updated