Rethink of impedance matching for traveling wave electro-optic modulators

Conventional design paradigms for traveling-wave electro-optic modulators (EOMs) traditionally require strict impedance matching. Yet practical ultra-broadband EOMs often employ terminal loads different from 50 Ω, suggesting this criterion warrants re-evaluation. This work revisits the impedance matching principles for traveling-wave EOMs using equivalent transmission-line models and numerical simulations. It is shown that employing a slightly low terminal impedance purposefully suppresses near-DC voltage build-up, which flattens the overall electro-optic response and artificially extends the 3-dB bandwidth without degrading high-frequency performance. Conversely, utilizing a moderately low source impedance maximizes the actual modulation voltage swing across the electrodes, which is beneficial for achieving better modulation depth. Experimental validations on 6-mm thin-film lithium niobate (TFLN) Mach-Zehnder modulators (MZMs) reveal that integrating a 36.3 Ω on-chip terminal resistor expands the 3-dB bandwidth from 68 GHz (with a 50 Ω external load) to 107 GHz. Despite this apparent enhancement, OOK transmission performance remains unchanged up to 120 Gbaud, with negligible variation in symbol error rate (SER). These results indicate that classical conjugate-matching-based RF design rules do not directly translate to optimal EO bandwidth performance, and instead suggest a co-optimization framework for electrode impedance and RF driver design in ultra-broadband EOM systems.