Fig. 1. (a) Depiction of PN junction, (b) PIN with intrinsic region width w=0.33  μm (PIN2), and (c) PIN with w=1.05  μm (PIN3). (d) Effective refractive index changes of PN, (e) PIN2, and (f) PIN3 junctions versus the applied reverse DC bias voltage with respective MZM arm lengths of 2, 6, and 6 mm. Dots are the experimental measurements, and lines correspond to the respective simulations of the whole modulation of the DC Kerr effect and carrier modulations.

Fig. 2. (a) Schematic view of the experimental setup used to measure the EOM from the MZM. DC voltage is applied to both arms; RF is either applied in single-drive or push–pull configuration (EDFA, erbium-doped fiber amplifier). (b) Effective index variations measured in push–pull configuration versus the reverse DC bias for a fixed RF peak amplitude of 1.4 V, (c) versus the RF amplitude for three reverse DC biases.

Fig. 3. Dots and the lines represent, respectively, the measurements and the corresponding fit or simulations. (a) Optical MZM transfer function for three electrical spectral components excluding intrinsic losses with P0 the maximum output power, PDC the static power, PΩ the modulation power at angular frequency Ω, and P2Ω at frequency 2Ω for the PIN3 junction by applying reverse VDC=6  V, VRF=2.0  V. (b) Amplitude of the refractive index modulation at angular frequency 2Ω versus the applied voltage VRF at frequency Ω for reverse DC biases from 0 to 15 V. Whatever the value VDC, it induces no variation of Δn2Ω. (c) Respective relative contribution of index variation in the Ω component from EFI linear EOM and from carrier modulation versus the applied reverse DC bias voltage.

Fig. 4. (a) Setup used to acquire eye diagrams (PPG, pulse pattern generator; DC, reverse DC bias; EDFA, erbium-doped fiber amplifier; DCA, digital communications analyzer). (b) Extinction ratio and signal-to-noise ratio at 10 Gbit/s by applying dual 4VppDATA/DATA‾ driving in push–pull versus the applied reverse DC bias. Eye diagrams for reverse DC bias of 2 and 30 V are embedded.

Fig. 5. Optical eye diagram display from 1 mm long PIN3 modulator by applying dual 2VppDATA/DATA‾ driving in push–pull and reverse VDC=30  V measured at (a) 80 Gbit/s and (b) 100 Gbit/s with a numerical six taps feed-forward equalization (FFE).

Fig. 6. Optical transmission of an unbalanced MZI for an applied reverse bias of 0 and 30 V. λr is the resonance wavelength, FSR(λr) the free spectral range, and Δλr the wavelength shift for bias voltage variation from 0 to 30 V.

Fig. 7. Measurements (dots) of PDC, the MZM output DC optical power; PΩ, the modulation at angular frequency Ω; P2Ω, the modulation at angular frequency 2Ω as a function of wavelength for (a) the 2 mm long PN junction using VDC=2  V, VRF=0.51  V, (b) the 5 mm long PIN2 junction using VDC=4  V, VRF=1.6  V, and (c) the 5 mm long PIN3 junction using VDC=6  V, VRF=2.0  V. The dashed lines represent the corresponding fit.

Fig. 8. Eye diagram, respectively, measured at 2, 10, 18, 30 V reverse DC bias at 10 Gbits/s using 4Vpp on each arm corresponding to ER 1.5, 2.4, 3.3, and 3.7 dB, and SNR 8.9, 12.8, 14.3, 15.6 dB.

Fig. 9. Optical eye diagram of 6 mm long PIN3 modulator measured at a data rate of (a) 32 Gbit/s and (b) 40 Gbit/s using 4Vpp on each arm and reverse VDC=30  V, with ER 2.7 dB and 2.3 dB, respectively.

Fig. 10. Cross section of the fabricated device viewed with an SEM.