Bachelor and Master Thesis

Below we list ideas for Bachelor and Master thesis. Please also view the Doctoral Thesis topics we offer, most of  which can also be scaled to Master thesis topics. We will be happy to explain in more detail and define a project that suits your interests and technical background. You could also work on any of these projects within an internship, to get to know us and the subject.

  1. Galaxy cluster velocities from kSZ: we will use the available Planck data for some well-known clusters and see what can be inferred on their cosmic velocities, under various assumptions on mass, temperature, dust content, etc
  2. Density inhomogeneity in galaxy clusters: we will use available SZ maps of two very well-known clusters (Coma and A2163) and combined the results with hard X-ray data from the NuSTAR satellite, to make a measurement of gas clumping using a completely new method!
  3. Mapping of a giant radio halo: we can choose one of the known clusters from our (tons of!) VLA data and use the existing VLA pipeline to try to make a radio halo map. 
  4. ALMA SZ image of a cluster: we will combine a large amount of ALMA Compact Array data for the El Gordo cluster (over 40h) and make a map, to see if there is any hint of a shock or other structures. We shall then do some model fitting in the uv-plane.
  5. Lensed CIB profile for a cluster: we will use a simple lensing model for a cluster for analytically predict how the flux amplification profile will look like from lensing the Cosmic Infrared Background (CIB).
  6. Determination of Hubble constant from joint SZ and X-ray modelling: We have initiated a project to utilize state-of-the-art X-ray data from the eROSITA satellite, and SZ data from the SPTPol telescope (and in the future from Simons Observatory), to model the value of Hubble constant with these two measurements. We look forward to implement several novel techniques for this task using these current and upcoming data.
  7. Separation of kinematic SZ and CMB-lensing with machine learning: Machine learning is a vast and rapidly growing tool in astronomical data analysis. In this project we aim to train a Convoluted Neural Network on simulated data of galaxy clusters, where the kinematic SZ and CMB-lensing signals are on top of each other, and explore how one can separate them while having only a handful of objects.
  8. Hydrogen Recombination Lines (H-RL). This project would pick up our long-term effort to harvest the ALMA archive for observations of mm-wavelength hydrogen recombination lines (H-RL). The aim is to determine the line fluxes and relate that to the infrared-flux, in order to establish H-RLs as reliable tracers of the massive star formation rate. We have a unique software package that supports this. A related issue to follow up on is the reliability of mid-infrared imaging photometry as a proxy for the star formation rate.
  9. High-redshift galaxy survey. In preparation for the FYST/CCAT-prime first year deep photometric high-redshift galaxy survey we want to develop innovative methods for source identification and source population analysis. Specifically, we will explore three novel approaches that can be explored in several Bachelor or Master theses. (i) Adapt and refine developments to exploit Herschel observations (Shirley et al., MNRAS 507, 129, 2021) through a Bayesian cross-matching analysis with informed priors, e.g., source position, redshift, and galaxy type from optical to IR identifications. (ii) A cross-matching analysis based on machine learning methods (An et al. 2019), trained on deep optical, Herschel, SCUBA-2 and ALMA observations of the COSMOS and GOODS fields. (iii) A holistic Bayesian fit to a parametric model describing the source ensemble properties, including redshift-dependent luminosity functions, SEDs, and biasing. We will test these analysis methods on simulated sub-mm imaging that includes all instrumental and sky noise terms. Such analysis will also inform on the best observing strategy, e. g. to balance map width vs. depth, and on the required relative depth in the observed photometric bands.
  10. LEGO VLBI. To illustrate the principle of radio interferometry we built a model of the ALMA interferometer from LEGO pieces. The telescope configuration can be modified and the uv-coverage is updates instantly to show the resulting image of a given target. We plan to build a similar LEGO model to illustrate VLBI, with a 1-meter large globe on which telescopes can be mounted. A possible thesis would involve to modify the necessary software and design model targets, such as black holes.