Augmented Portable X-ray Fluorescence Technology for Application to Soils: Towards Modular Soil-Specific pXRF Instrumentation for SOC Quantification

Abstract

<p>Soil organic matter and organic carbon are variables of critical environmental importance in terms of soil productivity and function, global food security, and climate change mitigation. The global population is projected to reach approximately 11 billion by the year 2100 which will necessitate increased global food production output. United Nations’ multi-models also predict atmospheric carbon dioxide (CO₂) concentrations to reach between 794 and 1142 ppm by the year 2100 which would exacerbate climate change, and is detrimental to society. In response to these problems, the United Nations has championed the 4p1000 initiative which aims to mitigate climate change via the sequestration of atmospheric carbon into agricultural soils (thereby potentially reducing atmospheric CO₂ concentrations, increasing crop productivity, and global food security). As a result, participating countries are attempting to sequester atmospheric carbon into agricultural soils necessitating advances in rapid and accurate soil organic carbon (SOC) quantification technologies to monitor these efforts.</p> <p>Portable X-ray fluorescence (pXRF) is a rapid, potentially high throughput, and cost-effective technique capable of elemental assessment. Portable X-ray fluorescence instruments are widely used to rapidly measure and map soil elements, however quantification of light elements that constitute the majority of organic materials is not yet possible with this technique (e.g. carbon). A novel modular sensor attachment for pXRF instrumentation was prototyped enabling concurrent volumetric SOC quantification during pXRF analyses. The novel sensor attachment was situated orthogonal to the analyzer surface area near the samples undergoing analyses (referred to as the “Z-Plane” sensor). The decay of X-rays in a sample is related to the bulk attenuation coefficient of the material and thus, radiation detected at the Z-Plane sensor is potentially relatable to sample organic content. This novel sensor attachment detected errant radiation near sample vessels during analyses and related those signals to SOM and SOC thereby extending pXRF capabilities as detailed in this thesis.</p> <p>This primary prototype was initially tested on soils (sand, Vertisol, Cambisol) spiked with SOM surrogates (sucrose, Lucerne), and produced quantitative soil specific simple linear regression (SLR) calibrations (R² 0.89-0.96) where the independent and dependent variable were sensor response and SOC via dry combusion respectively. The effects of sample depth and density on Z-Plane sensor response was assessed and methods were investigated to mitigate random variability in measurements associated with sample preparation and homogenization. Quantitative SOC data was generated on real-world woodland Nitisol samples using the novel Z-Plane sensor attachment (R 2 = 0.88, RSD = 19.09%, RPD = 2.67). The Z-Plane sensor, which cost approximately 100 U.S. dollars to build, produced SOC data that was comparable with existing spectroscopy instrumentation.</p> <p>The effects of geochemical variability in prevalent elements targeted by pXRF users was assessed using samples spiked with arsenic, zinc, and lead while employing different but commonly applied settings (detailed in the thesis), and using two pXRF instruments from different manufacturers (Bruker Tracer III SD and SciAps X250) (all R² > 0.71). A diverse group of different real-world soils (Vertisol, Luvisol, Ferralsol, Nitisol) from different land uses (uncultivated lands, farmlands, woodlands) were also investigated. An assortment of different, but common parameters and settings were applied and yielded varying results, from qualitative to quantitative (R² ≤ 0.90) supporting the potential of the Z-Plane sensor for quantifying SOC in parallel to pXRF analyses at a 30 second scan time. The sensor had difficulty when used in soils with low SOC content (approximately 0.3-0.5%) and future research vectors were suggested that may potentially increase efficacy and performance of this primary prototype including methods to better homogenize samples, and integration of more advanced components. </p>