Fig. 1. (a) SEM image and refractive indices at 1580 nm of the layer stack. (b) GeSbS waveguide structure with a superposition of the fundamental TE-mode profile. (c) Linear optical transmission of the 1 cm long chalcogenide waveguide.

Fig. 2. (a) Bidirectional nonlinear transmission setup. Following the injection from both chalcogenide facets A and B, (b) measured output spectra varying the average input powers Pin between 0.5 and 11 mW, and (c) experimental (solid lines) and simulated (dashed lines) output spectra at Pin=11  mW, where the nonlinear phase shift for the simulated spectra coincides to ϕNL(A)=0.59  rad and ϕNL(B)=0.39  rad.

Fig. 3. Injecting from (a) the facet A and (b) the facet B output spectral r.m.s. width 2σ measured as a function of the average input power Pin (open circles) and calculated for various nonlinear phase shift ϕNL (solid lines). Insets: extraction of the nonlinear phase shift as a function of Pin.

Fig. 4. (a) Experimental spectra for Pin=10  mW and (b) simulated spectra for ϕNL=0.54  rad by varying the second-order dispersion coefficient ϕ(2). (c) D-Scan traces showing the measured (open circles) and simulated (solid lines) variation of the r.m.s spectral linewidth 2σ with ϕ(2).

Fig. 5. (a) Output spectral r.m.s. width 2σ measured as a function of the average input power Pin (open circles) and calculated for various nonlinear phase shifts ϕNL (solid lines). Inset: Extraction of the nonlinear phase shift as a function of Pin. (b) Output spectra measured for Pin varied between 4 and 15 mW, with a top-hat-like spectrum measured at low power (linear transmission). (c) Experimental (open circles) and linear fit (solid line) of the ratio Pin/Pout versus Pin.

Table 1. Experimental Values at λ=1580  nm