MSc thesis: Characterization and optimization of the phase extraction algorithm for Deep Frequency Modulation Interferometry

Job Offer from December 15, 2021


Laser interferometry is a powerful technology for measuring tiny displacements as variations of optical path lengths. When the optical path lengths cannot be kept constant to within a small fraction of a wavelength, several techniques exist to increase the working range of the sensor over multiple fringes. Heterodyne interferometry is the “Rolls Royce” technique for large dynamic range displacement sensing. However, this technique involves complex optics and electronics, and it is generally bulky. As we aim to build multi-channel sensors that probe several degrees of freedom of a mechanical system, the need arises to develop interferometry techniques that allow for simpler optical and electronical setups.

One such technique is Deep Frequency Modulation Interferometry (DFMI), where an unequal arm-length interferometer is injected with a frequency-modulated laser, resulting in a photocurrent that varies at the modulation frequency and its higher harmonics. The photocurrent is converted into a voltage and digitized, and a computer algorithm, known as the DFM Software Phasemeter (DFMSWPM), computes the single-bin Fourier transform of the signal for the first few harmonics of the modulation frequency, and performs a non-linear fit of these complex amplitudes to a known model, which allows the extraction of the interferometric phase in real time.

The project

For this Master's project, the candidate will investigate the DFMSWPM algorithm, and determine its performance and limitations using computer-simulated signals. The robustness of the algorithm will be explored in a large parameter space, including physical parameters (e.g., modulation frequency, modulation amplitude, interferometer arm-length difference), analog signal quality parameters (e.g., amplitude noise, contrast), digitization parameters (e.g. sampling frequency, downsampling factor), as well as parameters intrinsic to the algorithm. The lessons learned from this investigation will then be used to improve the DFMSWPM algorithm, and enhance the readout performance of a real DFMI experiment.

Candidates should have

  • Good knowledge of the basics of the C programming language
  • Good knowledge of Python for data analysis
  • A good mathematical background


Dr. Miguel Dovale

Junior Scientist/Postdoc
Go to Editor View