Low-temperature plasmas (LTPs) at atmospheric pressure hold great promise for disinfection and sterilization applications. When compared to traditional sterilization technologies like autoclaving, LTPs may offer several benefits, including reduced energy consumption, lower operational temperature, absence of high pressure or vacuum requirements, shorter treatment times, and the absence of persistent hazardous compounds. Nevertheless, the understanding of the physics and chemistry of LTPs is still incomplete due to the numerous variables at play, particularly in air at atmospheric pressure. The aim of this dissertation is to improve the comprehension of the mechanisms responsible for bacterial inactivation in atmospheric pressure LTPs by conducting a thorough physical, chemical, and biological characterization. Special attention is given to performing measurements under conditions identical to those of biological treatments, addressing the often overlooked influence of the biological target on the plasma discharge. To examine the effects of indirect plasma treatments on \emph{E. coli} an atmospheric pressure surface dielectric barrier discharge (SDBD) plasma in air, powered by a nanosecond pulse generator, is employed. The treatments demonstrate bacterial inactivation up to 4-log reductions after 10 minutes of plasma exposure. In-situ FTIR spectroscopy reveals the presence of O$_3$, NO$_2$, N$_2$O, and N$_2$O$_5$, while laser-induced fluorescence (LIF), employing a picosecond laser, is used to measure the kinetics of NO produced in the plasma on a two-dimensional area in front of the DBD surface. The results show a correlation between the concentration of reactive oxygen and nitrogen species (RONS) with the relative humidity (RH) and with the plasma discharge power, measured using Lissajous figures. The results suggest that NO is not a main factor contributing to the inactivation of \emph{E. coli} in the plasma treatments examined in this study. To investigate direct plasma treatments, \emph{Bacillus subtilis} spores on monolayer membranes are treated using a nanosecond volume DBD (VDBD) plasma at atmospheric pressure in humid air, reaching complete inactivation with >5-log reductions after 1 minute of plasma exposure. The membranes are treated in the VDBD plasma discharge on the ground side and on the high-voltage side, showing no difference in the treatment results. Both in-situ FTIR and NO LIF are performed on this setup, showing the presence of O$_3$, N$_2$O, and $\sim$1 ppm of NO temporally decaying in between plasma discharges. To evaluate the impact of the electric field in direct plasma treatments, measurements of electric field induced second harmonic (EFISH) generation using a picosecond and a nanosecond laser are performed. The ps EFISH measurements, owing to the improved temporal and spatial resolution, reveal features previously undetected by the ns EFISH measurements, including a different electric field evolution depending on the position in the plasma discharge. To understand the mechanisms of the nanosecond plasma breakdown in humid air, a comparison with a simplified kinetic model has been carried out. The results provide insights on the ion and electron kinetics, revealing differences with previously studied nitrogen plasmas and highlighting the importance of electronegative species in the breakdown dynamics. Finally, preliminary tests for achieving sterilization standards are performed.