Both teflon and quartz PM2.5 filters collected from January to July 2021 at the monitoring site of the Department of Mathematics and Physics of the University of Salento in Lecce (Italy) were analyzed by integrating different characterization techniques (Particle Induced X-ray Emission PIXE, Isotope Ratio Mass Spectrometry IRMS, and Accelerator Mass Spectrometry AMS) at the CEDAD (Center of Applied Physics, Dating and Diagnostics) of the Department of Mathematics and Physics, University of Salento. The PM2.5 concentration analyses allowed to identify the variation of the main PM2.5 characteristics as a function of the season and the day of the week. This last characterization was integrated by the results from the PIXE, which allowed to identify the heavy elements and their concentrations. The main results showed the presence of different elements, such as S and Zn (considered as markers of anthropogenic sources for PM2.5) and Ca and Fe (as markers of natural sources). The concentrations of these elements showed a significant decrease during the weekend, mostly in the case of elements of anthropogenic origin, according to the data on the PM2.5 temporal evolution. Using the isotopic markers of carbon and nitrogen by means of the IRMS, we determined values of δ15N between 4.5 and 10.6‰, which are consistent with the origin of PM2.5 from anthropic combustion processes and a secondary contribution from vehicular traffic. Similarly, the values of δ13C obtained by IRMS were in the range between −24.4 and −26.7‰, generally associated with biomass combustion and with vehicular traffic. An analysis of the fossil and modern contribution was carried out on the PM2.5 filters by measuring radiocarbon using the integrated IRMS-EA system connected with the TANDETRON accelerator and AMS spectrometer. In more detail, we found a percentage of modern carbon in the range 71.6–92.4% that indicates a larger bio-derived contribution with respect to the contribution from fossil sources during the analyzed period. The parameters obtained from PIXE, IRMS, and AMS techniques were finally used as input for different ordination methods that allowed their deeper characterization.