The accuracy of composition measurements by atom probe tomography is often dependent on the selected operation mode as well as the applied measurement parameters. The detected hit characteristics, distinguishing between single and multiple events, along with the electric field, are also affected by parameter selection. In this study, atom probe tomography experiments were performed on a stoichiometric TiN coating in voltage as well as in laser-assisted mode with systematically varied laser pulse energies. The observed elemental compositions were compared with complementary ion beam analysis measurements. The influence of multiple detection events was investigated by two approaches: I) A modified local electrode served as a hardware filter, reducing multiple detection events from 78.8 % to 41.9 % and from 40.9 % to 5.6 % using voltage mode and laser-assisted APT (0.6 nJ), respectively, and II) unfiltered datasets were analyzed by data post processing. The latter allowed the study of ion species, particularly of emerging complex molecular ions associated with dissociation processes. Additionally, average electric fields were estimated and spatial considerations were made to investigate the evolution of charge state ratios and hit characteristics during the measurement. Filtering the measurements significantly improved the elemental accuracy. In voltage mode, hardware and software filtering reduced the discrepancy between reference and observed composition from 3.8 at.% to 2.1 at.% and 0.1 at.% within uncertainty limits. In laser-assisted mode, higher laser pulse energy increased the difference between unfiltered data and the reference composition, from 1.4 at.% (0.1 nJ) to 8.1 at.% (2.0 nJ). Ion species analysis of the datasets shows an increasing presence of complex ions (Ti2N) with raising laser pulse energy. Electric field studies reveal a decline from 40 V/nm in voltage mode to 36 V/nm applying a high laser pulse energy of 2.0 nJ, indicating insufficient field strength for neutral nitrogen re-ionization.
Read full abstract