A study of elemental composition of PM2.5 samples collected at the main campus of the Universidad Nacional Autónoma de México, comprising data from 2015, 2016, and 2019, is described. X-ray fluorescence analysis was employed to measure concentrations of up to 19 elements, complemented with ion chromatography for eight ionic species (for samples of 2015 and 2016), and thermo-optical analysis for organic and elemental carbon (for samples of 2015). Positive matrix factorization (PMF) was applied to develop receptor models for each year, to identify polluting sources and their contributions to total mass. Five sources were identified in every case. With the aid of PMF, the influence of biomass burning on a major episode in May, 2019, was recognized. Comparisons with previous studies in this or nearby sites are also presented. It should be noted that only limited compositional data in this site are available since 2005 for the aerosol fine fraction.