Nuclear PDFs

Abstract

We summarize the current status of global analyses of model-independent nuclear parton distri- bution functions (nPDFs). A carefully conducted global analysis of nPDFs in the era of heavy-ion collisions at the LHC is key to gaining access to new information through the data measured by the collider experiments. This implies incorporating as large and diverse a dataset as possible, as well as the latest theoretical calculations, in order to capture a faithful representation of nuclear structure that reflects the current state of knowledge. On the other hand, the uncertainties in par- ticular must reflect what is not known. For many observables, nuclear uncertainties represent the most significant source of error in current theoretical models; nevertheless, they should not, where possible, be artificially constrained (for example, through parameterization bias). The quality of a global (n)PDF fit is often measured by the $χ^2$ per degree of freedom, which is approximately one for a fit that describes the data precisely without overfitting. Many additional aspects allow to characterize a good fit, e.g. the $χ^2$ distribution per experiment, the shape of the uncertainties, and the sensitivity to different types of data. We inspect various fit quality indicators for fits obtained by different groups. Morever we pay close attention to the $A$-dependence. Traditionally, the A-dependence had a primarily functional role in global nPDF analyses and comparatively less attention has been paid to it when characterizing a fit. However, we have now entered the era of nuclei collisions at the LHC and need to investigate the $A$-dependence properly.