Comparing portable x-ray fluorescence spectroscopy instrumentation for metallome analysis of herbarium specimens

Abstract

The use of x-ray fluorescence (XRF) instruments for metallome analysis of herbarium specimens to discover hyperaccumulator plant species has gained popularity, but a growing concern arises about intercomparability from the use of different instrument makes and models. Therefore, this study aimed to assess the performance and comparability of the results generated by three different XRF instruments and three different quantification methods (empirical calibration based on XRF versus inductively coupled plasma atomic emission spectroscopy [ICP-AES] regression, in-built manufacturer algorithms, and an independent GeoPIXE software pipeline based on Fundamental Parameters). Three instruments with distinct specifications were chosen to improve the generalizability of the results, ensuring relevance to a wide range of instruments that may be used in the future for metallome analysis of herbarium specimens. Each instrument was used to scan a representative set of dried hyperaccumulator plant leaf samples, and their accuracy in quantifying elemental concentrations was then compared. The manufacturer algorithms overestimate the elemental concentrations and have the highest errors. The empirical calibrations have the closest mean concentration to the mean concentrations reported by ICP-AES, but can produce negative values. The independent pipeline performance is marginally better than the empirical calibration, but it takes substantially more time and effort to setup the Fundamental Parameters through reverse engineering the instrument hardware parameters. Using the GeoPIXE independent pipeline to extract the XRF peak intensity to use in the empirical calibration performs better than manufacturer algorithms, while avoiding the complicated setup requirements, and this should be considered for further development.