Research Review: LED droop: Lumileds strengthens Auger case
Researchers Aurélien David and Michael Grundmann determined that droop is correlated to a shortening of the nonradiative lifetime after studying a 430 nm GaN LED featuring a double heterostructure with a 15 nm thick InGaN layer. “The nonradiative lifetime is quantitatively compatible with Auger scattering, which supports Auger as being this non-radiative mechanism,” explains David.
Although the data presented by Lumileds is in quantitative agreement with Auger scattering, it is still possible that droop is caused by another mechanism. The origin of droop is highly controversial, and many different theories for its cause have been put forward over the last few years.
One alternative explanation for droop is interband absorption, which, like Auger scattering, is a non-radiative process that depends on a cubic power law related to the carrier density. However, David and Grundmann argue that interband absorption is related to optical re-absorption, and is strongly influenced by geometric factors. Droop, however, appears to have a universal, geometry-independent behavior.
Another explanation for droop is based on energetic carriers flying over the active region. However, the Lumileds’ researchers say that this can now be ruled out, because this process would not influence the lifetime of the carriers trapped in the active region. They have shown that this lifetime shortens at higher current densities.
One scenario that they cannot rule out is carrier capture and escape. However, they believe that this is unlikely, because there is no reason why this process should depend on the cube of the carrier density.
Probing the sample
Lumileds’ latest work on droop has focused on determining the evolution of radiative and non-radiative lifetimes by measuring the differential carrier lifetimes in LEDs, and combining this data with an internal quantum efficiency measurement that gave a peak value of 65 percent.
To determine differential carrier lifetimes, the researchers fabricated vertically injected LEDs, and drove them with 3 μs pulses to avoid device heating. A small, AC voltage was superimposed onto this series of pulses, and this provided a probe for the lifetime measurement.
By assuming perfect injection efficiency, the researchers calculated carrier density, and then the radiative and non-radiative lifetimes.
They found that the non-radiative contribution to the lifetime decreases as current increases, indicating the onset of an additional non-radiative process. At high current densities the radiative lifetime saturates, due to a process known as phase-space filling.
The researchers then compared their data with the standard model for the recombination rate, which involves the sum of: Shockley-Read-Hall recombination; radiative, bimolecular recombination; and an Auger process that depends on the cube of the carrier density. This model provided an excellent fit to the experimental data, and produced a value for the Auger coefficient of 10-29 cm6 s-1.
A. David et al. Appl. Phys. Lett. 96 103504 (2010)
