Injecting carbon dioxide into geological formations for long-term storage is considered integral to reducing greenhouse gas emissions. Residual trapping of CO2 is a primary storage mechanism, whereby CO2 ganglia are trapped in the pore space by capillary forces. Experimental knowledge of residual trapping processes in rocks is critical to the development of safe storage strategies. Here we present a quantitative low field 1H nuclear magnetic resonance (NMR) core flooding study of CO2 residual trapping in three different sandstones. It was found that transverse relaxation (T2) measurements were sensitive to the dissolution of paramagnetic ions from rock matrix minerals after exposure to carbonic acid; this response was observed on time scales relevant to core flooding experiments (i.e., minutes to hours). Subsequently, a brine aging protocol was designed and implemented to minimize this chemical effect, and hence, by applying the well-known T2-pore size relationship, changes in T2 time distributions during core flooding could be related to displacement of brine from pores of different sizes. Comparison of T2 distributions for partially CO2/brine-saturated cores to previously published data for N2/H2O systems shows an increased displacement of brine from small pores by CO2. Furthermore, results from cyclical brine/CO2 injections showed an increase in the total volume of residually trapped CO2 and an increase in trapping efficiency; compared to the results observed for N2/H2O, however, the improvement in trapping efficiency with cyclic injection was less pronounced. Potential causes for the observed differences are discussed in the context of effective N2 and CO2 wetting.