1. In earlier works (2004_PSSb-214_2668 and 2005_PSSc-002_2928) there is
comparison of simulated I-V curves with some experimental data. The
computations were made by SiLENSe, assuming a certain series resistance
of the LEDs. The most strong discrepancy is observed at low current
densities where tunnelling current is important. IQE as a function of
current has been calculated without accounting for Auger recombination.
That is why, there is no efficiency droop in the theoretical
characteristics.

2. In 2008_PSSc-005_2066 we have suggested that Auger recombination is
responcible for the efficiency droop, independently of Lumileds. Since
that time we are succeeded in predicting realistic IQE behaviour versus
current density. The analysis of the droop in various LED structure is
given in 2009_PSSc006_S804.

3. Detailed analysis of the I-V curve and Vf cannot be performed by
using the SiLENSe simulator only. This is because the current spreading
in an LED die depends on the electrode configuration. For detailed
analysis of the above issues a combination of SiLENSe and SpeCLED is
necessary (that is why we should assume a certain series resistance of
the LED to calculate the I-V curve; SpeCLED calculates this
characteristic on the basis of 3D current spreading analysis). The
comparison of our theoretical predictions with experimental data are
presented in the next papers (2007_PSSc-004_0045 and
2008_SST-023_125023). Our recent paper (2010_ThinGaN) demonstrates how
RATRO module is used to find a relationship between IQE and EQE and how
the LED chip design can be improved by modelling.