Ranging Atom Probe Spectra for Multi-Element Analyses

Atom probe tomography (APT) is becoming an essential nano-characterization tool in a wide variety of fields. In principle, APT can deliver 3-dimensional chemically resolved images of nano-scale analysis volumes, thereby making the technique uniquely suited for applications requiring composition measurements at dimensions approaching atomic-scale (e.g., semiconductors) and nano-scale analysis in fields such as nano-geology. In practice, APT analyses can be hindered by two broad types of limitations – ion trajectory aberrations and composition measurement bias – leading to image distortions and erroneous chemical analysis results. However, these limitations are not constrained by physics, and thus should be possible to overcome through measurement science, technological advances, and application of standards. Ion ranging – i.e., how to properly attribute regions of interest in the spectrum to specific ion species – and associated measurement bias have long been topics of discussion within the APT community. APT spectra are complex, with peak forms often varying significantly between ion species and between data sets. No single model peak form can be applied universally to range spectra and there is no community-wide consensus on how best to range spectra, which can lead to erroneous analysis results and lack of reproducibility.<br/>We have previously achieved accurate and repeatable single-element isotopic analysis results by employing a three-step analysis approach that relies on data filtering (artifact removal), adaptive peak fitting (assumes all isotopic variants within a given ion species have the same peak shape), and standards-based calibration (using known reference materials) [1-3]. However, this three-step approach cannot be used for complex multi-element analyses. In multi-element analyses, the peak forms often cannot be assumed to be identical, as required for adaptive peak fitting, and filtering the data to remove artifacts can introduce additional composition measurement bias. An alternative methodology must be developed.<br/>We present the results of a systematic ion ranging study for multi-element analyses that compares about ten different APT peak sampling methodologies. A natural mineral specimen was chosen for the study since the material was characterized by multiple complementary methods and produced complex multi-element APT spectra with a high fraction of multi-hit detection events. Clear trends in the analysis results were observed for the measured element ratios of interest. The trends could be demonstrably associated with peak sampling bias and data interpretation artifacts. By applying a custom correction method to account for peak sampling bias, the analysis results improved significantly, with up to a 10-fold reduction in bias for the relevant peak ratios measured. Surprisingly, all APT ion ranging methodologies explored yielded similar analysis results, once peak sampling bias was accounted for. Further, the APT results largely agreed with the other complementary analysis methods employed for validation.<br/><b>References</b><br/>[1] P. Gopon, J.O. Douglas, F. Meisenkothen, J. Singh, A.J. London, M.P. Moody, Atom Probe Tomography for Isotopic Analysis: Development of the 34S/32S System in Sulfides, Microsc. Microanal., (2021) 1-14.<br/>[2] F. Meisenkothen, M. McLean, I. Kalish, D.V. Samarov, E.B. Steel, Atom Probe Mass Spectrometry of Uranium Isotopic Reference Materials, Analytical Chemistry, 92 (2020) 11388-11395.<br/>[3] F. Meisenkothen, D.V. Samarov, I. Kalish, E.B. Steel, Exploring the accuracy of isotopic analyses in atom probe mass spectrometry, Ultramicroscopy, 216 (2020).