Nils Schöneberg

One of the most pressing issues for modern cosmology is the Hubble tension between the measurement of the Hubble parameter based on the local distance ladder using Cepheids and the inference using the CMB measured from Planck in the context of the ΛCDM cosmological model. While there are a lot of alternative measurements as alluded to in my talks ,extensive cross-checks are still essential, especially as most probes don't reach quite the same precision. For this purpose, I have investigated constraints from a combination of Baryonic Acoustic Oscillations (BAO) and Big Bang Nucleosynthesis (BBN) that provide alternative constraints to H₀ than those from the CMB, in my first and second paper on this topic. In particular, in these works we independently confirmed and extended previous work using our own updated BAO and BBN likelihoods. We checked the dependence on the BBN data and theory as well as various combinations of the newest BAO data from BOSS/eBOSS together with other clustering data, giving us one of the tightest constraints on the Hubble parameter independent of the CMB to date. We also show that this combination of probes depends on the assumed underlying cosmological model, especially if that cannot be constrained from BBN or BAO.

In my future research I would like to tackle this tension, as well as the S₈ tension in a more model-independent way.

The nature of dark matter is currently one of the biggest unsolved puzzles of modern particle physics and cosmology. Whereas we are confident about certain properties of dark matter (clustering strength, lack of very strong interactions, current abundance), there are other properties which are almost unconstrained. This can be both a great opportunity for creative model building to reconcile disparaging measurements but also a stifling lack of clear direction in the vast model space. As such, my research focuses on models of dark matter and dark radiation which are phenomenologically and theoretically well motivated and at the same time remain falsifiable within the next decades. In my current research I have investigated cannibalistic dark matter, multi-interacting dark matter, Wess Zumino Dark radiation, and the Bekenstein model. While these individual investigations are certainly necessary and provide deeper understanding of the individual models, a unified approach of investigation is rather crucial. This is what the widely known H₀ olympics paper is about.

In my future research I would like to further qualify the most important phenomenological aspects of different dark matter models that allow them to ease existing tensions and how to constrain these in a model-independent approach.

The evolution of the thermal state of the universe throughout cosmic history is one of the most important pillars of modern cosmology. Understanding it is crucial for many different probes from BBN over the CMB to the Lyman-α forest. In my research, I have worked on the question of how a future detection of Spectral Distortions of the CMB could aid in constraining the ΛCDM cosmological model and exotic energy injections by implementing a consistent treatment of energy injection and the numerical evolution of the thermal bath temperature within the CLASS code, alongside the implementation of a spectral distortion module and a python likelihood including astrophysical foregrounds together with my collaborators. I have also worked on the use of the Lyman-α forest as a valuable cosmological probe for the content of the universe, putting tight bounds on neutrinos, thermal warm dark matter, and interacting dark matter in my first and second paper using the Lyman-α forest.

While the previous aspects have focused on specific models and measurements to investigate, this aspect is more focused on how to investigate these. With the advent of increasingly large data sets derived from cosmological experiments and the multitude of probes to be considered, the importance of fast and accurate calculations of cosmological observables of all kinds cannot be understated. Here, my research has branched across several different ingredients. In detail, I have implemented an alternative way of calculating angular power spectra beyond the traditional line-of-sight approach (see here), which allows for a speedup of the computation of these spectra by several orders of magnitude. This method has also recently been submitted to the N5K challange of the DESC for LSST. Furthermore, I am collaborating on a project (titled CosmicNet) to use machine learning predictions to replace lengthy calculations of differential equations based on analytical understanding of the functions to be predicted (first and second paper). Finally, I am also working on a project involving Gaussian Processes for fast likelihood emulation.