How do I use a photonic link for frequency conversion of RF signals?

Photonic frequency conversion uses the nonlinear mixing properties of electro-optic modulators and photodetectors to translate RF signals from one frequency to another entirely in the optical domain: (1) Method 1: modulator mixing: apply the RF signal to one MZM and the local oscillator (LO) to a second MZM. The optical outputs of both MZMs are combined and detected by a photodetector. The PD output contains mixing products at: f_RF ± f_LO (just like an electronic mixer). The desired product (upconverted or downconverted) is selected by an RF bandpass filter. (2) Method 2: single-modulator mixing: apply both the RF signal and the LO to the same electro-optic modulator (superimposed on the modulator drive). The modulator transfer function creates intermodulation products, including f_RF ± f_LO. This is simpler but has lower conversion efficiency and more spurious products. (3) Method 3: heterodyne detection: modulate the RF signal onto a laser at wavelength λ₁. Use a second laser at wavelength λ₂, offset by the desired LO frequency: f_LO = c × |1/λ₁ - 1/λ₂|. Combine both optical signals on the photodetector. The PD output contains the beat frequency: f_RF ± f_LO. This requires the two lasers to be phase-coherent (locked to a common reference or each other). (4) Advantages over electronic mixers: wideband: photonic frequency conversion works from DC to 100+ GHz (limited by the modulator and PD bandwidth, not by the mixer design). No image frequency problem: the optical carrier separates the signal and image (in some architectures). Tunable: the LO frequency is easily changed by tuning the second laser (for heterodyne) or the RF LO (for modulator mixing). EMI-free: the mixing occurs in the optical domain; no RF LO leakage to worry about. (5) Disadvantages: conversion loss is typically higher than electronic mixers (20-40 dB for a simple photonic mixer vs 6-8 dB for an electronic mixer). Noise figure is higher (dominated by the photonic link NF). Cost and complexity are higher.