Thesis

Modern cosmology is currently facing major challenges. On the one hand, the standard cosmological model, known as LambdaCDM, relies on two components, dark energy and dark matter, whose effects are now well measured, but whose fundamental nature remains unknown. On the other hand, as observational precision continues to improve, increasingly significant tensions are emerging within the LambdaCDM framework. In particular, measurements of the BAO (Baryon Acoustic Oscillations) feature from the large-scale spectroscopic DESI survey provide a way to measure the expansion history of the Universe. When interpreted within the LambdaCDM framework, the results obtained so far are difficult to reconcile with measurements of the Cosmic Microwave Background (CMB). This tension could be caused, for example, by dynamical properties of dark energy or by other new physical phenomena. It has attracted widespread attention across the scientific community: over the past two years, DESI papers on this topic have ranked among the most highly cited publications across all fields of astrophysics and particle physics. To distinguish between the various hypotheses proposed to explain these findings, one approach stands out: obtaining new, independent, and more precise measurements.

The aim of this PhD project is precisely to obtain the most precise possible cosmological measurements at high redshift (z > 2), corresponding to the first few billion years of the Universe, in order to distinguish between different cosmological models. To achieve this, we will use the best probe of the matter distribution at these redshifts: the Lyman-alpha forest, a collection of absorption lines measured in the spectra of distant sources. These lines provide a genuine tomographic view of the early Universe at redshifts inaccessible to galaxy surveys, making them complementary to other cosmological probes. The PhD will be carried out directly within the DESI collaboration, which has already collected one million Lyman-alpha spectra, by far the largest sample of its kind available worldwide. By measuring the two-point correlation function of these data, in a manner analogous to galaxy clustering analyses, we can measure the BAO feature and infer cosmological distance scales. The DESI DR2 Lyman-alpha measurement has already achieved a combined precision of 0.7%. The aim of the PhD will be, first, to significantly improve this measurement by developing novel methods and, second, to exploit the information contained in the data across the full matter field, including scales beyond those associated with the BAO feature, to constrain cosmological models.

This PhD will be carried out within the cosmology group at IRFU/DPhP (CEA Paris-Saclay), which has been at the forefront of Lyman-alpha forest analyses for over a decade and has played a major role in the DESI collaboration since the project began. In particular, the student will be an active member of the DESI Lyman-alpha Working Group, co-led by Eric Armengaud from 2024 to 2026, and will work closely with our European and American partners within this group.

To extract the maximum amount of cosmological information from the data, the PhD project will focus on the following methodological innovations:
– Improving BAO precision. The student will develop, for the first time, a BAO reconstruction method for the Lyman-alpha forest. This method is already well established for galaxies but requires technical innovations to be applied to Lyman-alpha data. By partially correcting for the effects of gravitational motions that have blurred the BAO signal, it will reduce the uncertainty on the position of the BAO feature and refine measurements of the cosmic expansion rate. The method will first be validated using simulated data before being applied, as part of the PhD, to DESI DR3 and DR4 data.
– Exploiting all accessible scales of the matter density field for cosmology. Building on similar efforts within our group using DESI galaxy data, the student will implement a new method known as simulation-based inference (SBI). This approach uses numerical simulations to exploit the full matter density field, including its nonlinear regime. This work will pave the way for novel combined analyses and multi-probe constraints on the growth rate of cosmic structures and on primordial non-Gaussianities, which provide a signature of inflation.

These two contributions will directly support the delivery of DESI’s final cosmological results using the Lyman-alpha forest — an important milestone at a time when the global cosmology community is questioning the foundations of its standard model. By making optimal use of a dataset unparalleled worldwide and pioneering next-generation analysis methods, this PhD project will deliver some of the most competitive constraints available on the nature of dark energy and help determine whether the current tensions are evidence of new physics.

This PhD project would preferably be preceded by a research internship within the group.