Regensburg 2019 – wissenschaftliches Programm
HL 3.7: Vortrag
Montag, 1. April 2019, 11:15–11:30, H33
Fabrication of spectrally homogeneous microlaser arrays as a nanophotonic hardware for reservoir computing — •Tobias Heuser1, Jan Grosse1, James Lott1, Daniel Brunner2, Ingo Fischer3, and Stephan Reitzenstein1 — 1Institut für Festkörperphysik, Technische Universiät Berlin, D-10623 Berlin, Germany — 2FEMTO-ST, 15B Avenue des Montboucons, 25030 Besançon, France — 3IFISC (UIB-CSIC), E-07122 Palma de Mallorca, Spain
Reservoir computing is a powerful machine learning concept for a new kind of
data processing which is inspired by the neurons in the brain. In this concept an interacting network of nodes is evaluated by a trained readout for applications like fast pattern recognition. To further improve the performance of this
concept, a photonic hardware implementation is of particular interest.
Here, we report on our new developements in the fabrication
process and lasing performance of large 2D arrays of microlasers, namely quantum dot micropillars and VCSELs. These arrays will serve as a nonlinear network via diffractive optical coupling . For this spectral alignment of the involved lasers is crucial. To achieve this with a spectral homogeneity better than 200µeV throughout the array of up to 900 lasers, shifts of the emission energy are compensated by electrical tuning or by precisely adjusting the radius of the fabricated micropillars [2, 3].
 D.Brunner, I.Fischer, Opt. Lett. 40, 3854-3857 (2015)
 S.Reitzenstein, A.Forchel, J.Phys.D.Appl.Phys. 43, 033001 (2010)
 T.Heuser et al., APL Photonics 3, 116103 (2018)