Novel Measurement Scheme for Injection Locking Experiments with Semiconductor Lasers - Lasers and Electro-Optics Europe, 1998. 1998 CLEO/Europe. Conference on

injection locking ex-snts with pemicondvctor 1-n Wed in the "rly eighties. A lot of w o k WBS performed since then mainly dcaliig with loeking range and stability, noise, modulation, chirp and linaidth nduetioo. sidemode injectim lackink optical frequency conversion and micmwave signal generation. However, there is a lack of quantitarivr comparison between experimental d m and theory. The main reason for h i m relies on the fact thsl the experimental ret-ups used in moa former studies are based on the -space pmpagation of light. This EBUISI difficulties in cantrolling the SXBFI amount of lighl injecrcd into the slave laser cavity, because the spetial overlap between the optical mode in free-space and the slave Iher waveguide is unknoun. The cxperimenml ret-up pmenfed in this paper allows to overcome this problem. The light generated by the master laser travels thmugh an optical fibre behre being injeded into the cavity ofthe slave laser. Uringthe mipmity principle, it S pssible to conml accUIBfely the absolute light power i n j d from the mastn IUET in10 the cavity of the slave laser mce the power coupling coefficimt berwren the slave waveguide mode and the fibre mode i s deremined. In addition, the polvisation ean be filly conbolled derpileths birefringence of the optical fibre. The master and d a w used in the experiment were both DFB diode lasm mining near 1.3 @m]. Vmim measurements were Farrid out rhowing typical injection locking behaviour with semicondustor lawr (e.g. stzble loeluog, undamped relaxation orillations, cham, nearly degenerate four-wave mixing and pciicd doubling). The messurements wen compsnd -fully with the mults fmm a standsrd mtKquatior bared model by means a f t h e correspdiog power spectra. Measurements and model were limked anal,tically to make a quantitative canparison possible. Knowing the power COLIpling between the optical mode in Le fibre aod the slave lam waveguide, it turned out ulal measurements and model could be related entirely via the distributed cavity loss ofthe slave laser. The spectra were nomalid with re~pect o the power speckurn of the frrcrunning slave laser. M e ~ ~ u r e d and FBIculatd power spfftrn agree very well in all nihlatims, fmm elaSsiCal injection locking to chaotic behaviour. Ditmcnces mainly slem fmm =me mismatch b e e n the arperimenlal and theoretical o p t i n g pine, inaccurate parameter vducs and 6om noix being neglected in the model. The pre-led m-up may be wily adaped fm experiments with passive optical feedback. Future work using this experimental set-up will be performed with diode laxrr emitting at 1.55 [bm], &BP Rayleigh bachueriog in the fibre and thus ucdesired light feedback into the slave Iis reducedsubstsntially.