What is the problem with the Auger recombination coefficient?
As many of you know, the Auger coefficient C is a problematic parameter. In the old days, when I was looking into the temperature sensitivity of long-wavelength laser diodes, reported C parameters varied over two orders of magnitude – for the same material ! Unfortunately, nobody found a way to measure the Auger coefficient directly. Experimental investigations are typically based on the modeling of measured light emission characteristics. The resulting C parameter strongly depends on the experimental conditions and on the employed model.
Now we face a similar problem with GaN-based short-wavelength emitters. In fact, the relevance of Auger recombination in those devices is often disputed. First direct experimental evidence of Auger recombination in InGaN/GaN LEDs was presented only recently, but it did not nail down the magnitude of the Auger recombination rate and its contribution to the much discussed LED efficiency droop. All quantitative assessments of Auger recombination are still based on modeling. Extracted C coefficients are widely scattered, as shown in the picture.
However, thanks to the intense worldwide effort to improve the GaN-LED efficiency, I hope that there finally is a chance to get to the bottom of the C parameter problem. We just published an investigation that successively applies more accurate Auger recombination models to the same LED efficiency measurement.  This study reveals the strong influence of quantum well properties on the extracted C coefficient. It demonstrates that the simple ABC model is unable to deliver a consistent Auger parameter and that bulk InGaN coefficients cannot be applied to quantum well devices. Comparable microscopic computations of quantum well Auger coefficients are not available yet, but I am quite optimistic that we will eventually be able to reach a consensus on this important parameter.
 J. Piprek, F. Roemer, and B. Witzigmann, Applied Physics Letters 106, 101101 (2015).