Comments on Zihang Zhu, Shanghong Zhao, Xuan Li, and Kun Qu, Tao Lin, “Photonic generation of frequency-octupled microwave signal with reduced electrical local oscillator power and improved spectrum purity” Opt. Quant. Electron (2017)

In the title paper, Zhu et al. presented a technique for generating filter less frequency octo-tupling signal based on a cascade of two dual-polarization modulators. Firstly, the harmonics of order 4\(l\) (\(l\) in an integer) are generated in the \(x\)-polarization by driving the x-modulators with an RF signal with a 90° phase difference between the two stages. Then the carrier is suppressed using unmodulated \(y\)-polarization by means of polarization multiplexing. Accordingly, the demerit of their design is the requirement to reset a precise polarizer angle to suppress the optical carrier every time there is a change in the RF drive amplitude or the modulation index. Nevertheless, the authors claim superior performance to a similar architecture in which the y-branch is modulated with the proper RF drive phases to suppress the carrier for all drive levels with no need to adjust a splitting ratio. The aim of this comment is to show that by modulating the \(y\)-polarization with adjusted RF drive phases, same architecture can be used to perform frequency octo-tupling whose performance is neither limited by precise settings of the polarizer angle nor by the proper settings of the modulation index. The carrier is suppressed for a unique value of the polarizer angle for all modulation index. If the modulation index is adjusted to suppress the \(4{\text{th}}\) harmonics, same circuit can be used to perform frequency 24-tupling generation.