Energy transfer between the interacting waves in a distributed Brillouin sensor can result in a distorted measurement of
the local Brillouin gain spectrum, leading to systematic error. We demonstrate here that this behavior can be fully and
precisely modeled, and an excellent quantitative agreement is found with experimental tests. Strict guidelines can be
enunciated from this description to make the impact of depletion negligible, for any type and any length of fiber.
The energy transfer between the two interacting optical waves in a distributed sensor based on stimulated Brillouin
scattering can lead to a non-uniform spectral distribution of the pumping power after a long propagation. This results in a
spectrally distorted gain that biases the determination of the maximum gain frequency. A quantitative analytical model
gives an expression for the tolerable pump power change keeping the maximum bias within a given accuracy.
A compact slot waveguide polarization splitter based on the silicon (Si) material system is proposed and analyzed. The slot waveguide structure introduces significant beat-length differences for transverse electric and transverse magnetic polarizations at operation wavelength of 1.55 µm. The special self-imaging design with limited modes and weak second-mode excitation shortens the device length to <50 µm, while still delivering good performance. The polarization extinction ratio is 13 dB, and the excess loss is only 0.3 dB. The splitter also has a relaxed length tolerance of 550 nm. These attributes make it an excellent candidate for polarization diversity circuits and compact electronic-photonic integrated circuits.