Galaxies are expected to grow and evolve via a series of physical processes relating to gas flows into and out of the galaxy. Inflows of gas from the surrounding cosmic web provide fuel for star formation, which subsequently causes an enrichment of the interstellar medium (ISM) with the metals produced within stars, whilst supernovae-driven outflows drive gas out of the galaxy, re-distributing metals in the process. In this way, measurements of chemical abundances within galaxies can provide insight into the different physical processes that drive galaxy evolution. The interplay between these different processes has been well studied in the local Universe by large spectroscopic surveys that have established a number of scaling relations between stellar mass, star formation rate and gas-phase metallicity. However, the existence of such relations at earlier times in the Universe is less well studied. The aim of this thesis is to investigate the evolution of chemical abundances within galaxies across cosmic time, making use of integral field spectroscopic data obtained through the KLEVER survey.

In the first part of this thesis, I compare the galaxy-integrated properties of galaxies at z=2 to those found in local galaxies, with a particular focus of the abundance of nitrogen relative to oxygen (N/O). I find that high redshift galaxies have similar N/O values to local galaxies at a fixed metallicity, but much lower N/O values than local galaxies at a fixed stellar mass. I then demonstrate that an anti-correlation exists locally between N/O and star formation rate, such that at a fixed stellar mass galaxies with higher star formation rates have lower N/O values. In light of this, I parameterise a three-dimensional relationship between stellar mass, star formation rate and N/O abundance, before demonstrating that this relationship accurately predicts the N/O ratios of galaxies at z=2 as well as those observed locally. As such, I name this relationship the fundamental nitrogen relation (FNR), in analogy to the fundamental metallicity relation (FMR). Furthermore, I show that the measured FNR is well described by a simple combination of the FMR and a non-evolving relationship between N/O and metallicity. These results suggest that the physical processes that govern the FMR must be sensitive not only to the metallicity, but also the N/O abundance.

In the second part of this thesis I extend my analysis to the spatially resolved scale, studying the spatial distribution of N/O in galaxies at z=2. I present some of the first measurements of N/O gradients at z=2, finding they are generally flatter than those found locally. This is contrary to inside-out growth models, which predict steeper gradients at earlier times, however this difference may be reconciled by invoking star-formation driven feedback mechanisms that effectively mix metals within the ISM. I present observations of inverted N/O gradients, which I suggest may be a consequence of the inverted metallicity gradients also observed at high redshift. I also present evidence for negative Balmer decrement gradients within z=2 galaxies, consistent with high levels of star formation in the galaxy centre that may be associated with early bulge formation. I note that the slope of the N/O gradients is dependent on the choice of diagnostic used to determine the N/O, suggesting this may be driven by differences in the ionisation properties of sulphur relative to oxygen.

Finally, in the third part of this thesis I present preliminary work analysing the scatter in the relationship between N/O and O/H for local galaxies. I present observations of a population of galaxies with low metallicities that have enhanced N/O abundances. I show that the galaxies with the highest N/O values also have higher stellar masses and star formation rates. I then investigate the possibility that these galaxies have undergone recent gas accretion, driving changes in their metallicities and N/O values whilst boosting their star formation. I compare to a simple gas mixing model, finding that the deviations of galaxies from their expected metallicities and N/O values can be well modelled by the accretion of metal rich gas with a metallicity equal to 55% of that of the galaxy. However, the models also predict that the gas fraction within the galaxy is expected to increase by between 0.64-1 dex during the accretion event, much larger than the changes in gas fraction inferred from the observed deviations from the star forming main sequence for local galaxies. I demonstrate that the expected changes in gas fraction are better matched by accretion of lower metallicity gas, however such models are unable to reproduce the observed decrease in N/O from the expected values. I conclude that improved models are needed that include prescriptions for star formation, chemical enrichment and gas outflows in order to better constrain the impact of dilution events on the N/O values and metallicities within galaxies.