GIRO, a giant reflectivity to zero order -grating

Research Area: III-V Semiconductor Photonic ICs

Main Researcher: Lieven Vanholme

The polarization of the light, emitted by a VCSEL (vertical cavity surface emitting laser) depends not only on the crystallographic direction. It'll change, due to temperature-, currentshifts and optical feedback. It's hard to control all these parameters and obtain stable polarized light. Is a fixed polarization that hard? No it isn't. A simple polarization stabilization structure can handle the job, e.g... the GIRO.

This grating-structure, on top of the laser, gives on certain circomstances GIant Reflectivity to zerO order. If the period of this grating has the same order as the entering wavelength, a funny resonance effect can make the grid to act like a mirror. The clue of the GIRO is that the polarisation of the light takes part in the game. The plot on the right shows the gratingheight(x)-, gratingperiod(y)- dependencies for the 0th order reflectivity on a GIRO (n=3.5). The huge difference between the 'almost 100% reflection'- tops en the 'near zero'- valleys is amazing!

It's pretty easy to calculate & simulate this reflection for flat waves falling into an infinite grating. But it isn't very physical. Using CAMFR we'll try a more realistic approach. With new simulations we want to examine the behaviour of a gauss-shaped wave falling into an finite grating. The goal is to find parameters for wich the 0th order reflection is still (very) heigh for just one polarization.

first three pictures shows a working GIRO , (height/wavelength = 0.74, period/wavelength = 0.66) reflecting 91,1 % of the 0th order (91,7% for gauss),last three are gratings with slightly dfferent period (height/wavelength = 0.74, period/wavelength = 0.60) reflecting 2,3%. (3,6% for gauss)

reflection (gratingheight=x)(periodsize=y)

By adding the giro on top of a VCSEL we can reduce the number of DBR pairs. The device get smaller, leading to a better mode confinement.

field profile VCSEL (oxide window)

Because of the reduced heigth of the laser, a smaller diameter will raise the minimum VCSEL gain less quick as it does for normal VCSELs. Making it possible to construct smaller Lasers with an acceptable low gain (e.g. < 500 cm-1).

gain

Other people involved:

• Lieven Vanholme

Related Research Projects

• IWT: gemini