Fig. 1. SEM image of the 125 µm fiber cross section. The anode hole is the one closest to the core.

Fig. 2. Intensity response of a Mach-Zehnder interferometer with the fiber component as the active arm. V_π~1.37 kV at 1550 nm after poling. The intensity response is nonlinear with applied voltage both before and after poling since Δϕ;∝(E_ext+E_rec)² The inset shows the parabolic relation between Δϕ and applied voltage, V_ext, obtained by fitting Eq. 1 to the measured intensity. The signal is noisier before poling since the fiber was not as firmly mounted as after poling.

Fig. 3. Plot of optical intensity versus applied voltage for the data points shown in Fig. 2 after poling. Two relatively good fittings are shown using Eq. 1, where very different values of Vrec are used (1200 V and 1950 V). This shows that the fit using such a limited phase excursion is very imprecise when determining the recorded field. Likewise, the value of χ ⁽³⁾ inferred varies from 1.86×10^-22 m²/V² (violet curve) to 2.36×10^-22 m²/V² (orange curve). Both fittings are well within the error margin of the experiment.

Fig. 4. Intensity modulation of Mach-Zehnder interferometer at 1.55 µm with poled fiber in active arm as the DC bias is changed. The minimum corresponds to the external voltage that cancels the effect of the recorded field.

Fig. 5. Poled 125 µm fiber etched for 45 seconds and imaged using a phase contrast microscope. On the right, simulation of the field applied during poling.