Spatial coherence of SLED light (width of active region?) As far as I know, super-luminescent diodes (a.k.a. SLEDs) have a very similar structure to laser diodes, except that the front faces of the active region are tilted to prevent the sort of reflections that would benefit lasing. However, I know that a laser diode below the lasing threshold emits multi-mode (i.e. spatially incoherent) light, as one would expect since the mode selection induced through the lasing is non-existent below the threshold and the active region is broad enough to theoretically allow for many transverse modes. SLEDs, on the other hand, are commonly claimed to have a near-single-mode output (i.e. high spatial coherence), which also seems evident from the fact that they are often effectively coupled into single-mode fibers. My question therefore is: 
Why do SLEDs predominantly emit light in the fundamental mode? Is this simply based on a narrower active region compared to laser diodes, which geometrically benefits lower modes (similar to how the small core of a single-mode fiber only allows the fundamental mode to occur)?
I tried to look up the width of the active region in a few SLEDs' data sheets, but it seems that such information is not commonly provided. Thanks in advance for any help.
 A: I believe the reason is indeed that the active region is small enough to only support one transverse mode. However, I have to disagree somewhat with part of your premise:

However, I know that a laser diode below the lasing threshold emits
  multi-mode (i.e. spatially incoherent) light...

This depends very heavily on the laser diode. I'm not sure what type of diodes you have experience with, but many laser diodes, including 100% of the ones we use in my lab, are "single mode" below threshold as well. They are small enough to only have a single mode and so don't need any mode suppression antics. For example, eagleyard specifically sell single-mode laser diodes and multi-mode laser diodes (which are capable of outputting much higher powers).
The single mode laser diodes from eagleyard output, as a reference, can easily output on the order of 100mW, which in my experience with superlum SLEDs is fairly consistent with the power you get from a single-mode SLED, so I see little reason to suspect anything more complicated than this. 
As you point out, it's seems hard to get hard numbers on the active regions for most of these things, and I didn't find a spec for these single-mode laser diodes. However, I was able to get a couple related numbers which support this picture. Eagleyard tapered amplifier (TA) lasers are intended for use as single-pass laser amplifiers, having a narrow input and a broad output to couple single mode light into a large gain medium and only amplifying the single mode. The input for a 850nm 2W TA is spec'd at 3um. I suspect that this size is picked to match the single-mode laser diode size, and is definitely small enough to only support one transverse mode, but it's hard to be positive. The output of this TA is spec'd at 210um, and similarly the active area for the eagleyard multimode diodes is 200um.
A: Maybe your confusion comes from the fact that a laterally single mode laser diode can still emit multi-mode light? This is because of longitudinal modes (because the optical laser length is much longer than the wavelength).
The difference of a LD and SLD maybe that a higher order transverse mode of a laser does not reach threshold and thus has a much lower intensity relative to the fundamental mode. SLDS do not have this thresholding effect and the suppression of higher order transverse modes maybe less than in a LD which then negatively affects the spatial coherence.
Note that the waveguide in transverse single mode lasers is never a truly single mode waveguide as in a single mode fiber. It is typically much wider. This is because the thresholding effect suppresses the higher order transverse mode (because it has higher loss). To reach the same transverse mode performance with an SLD it may need to be narrower than an LD.
A: Traditional light sources (candles, light bulbs, star light) can be just as spatially coherent as the best laser.  After all didn't Thomas Young perform his famous Young's double slit experiment in the early 1800s.  
In your case, the the SlED has a single mode waveguide that acts like a spatial filter to only propagate certain modes of light.
