Development and optimization of microbeam X-ray fluorescence analysis of Sr in speleothems

Abstract

Trace-metal abundances and ratios (Sr/Ca, Mg/Ca, Ba/Ca) in speleothems can be indicators of the hydrogeochemical processes active in overlying epikarst and serve as valuable proxies for precipitation amount and source, water residence time, and vegetation cover. However, conventional methods of trace element acquisition can be expensive, time-consuming, and destructive. A semi-portable µ-XRF system is compared to other methods, as it is non-destructive, rapid, and has the capacity to provide high quality datasets. A suite of test scans was performed in order to define the appropriate methodology for the application of µ-XRF in trace metal analysis of speleothems. Statistical analyses were performed on an experimental dataset in order to assess the precision with which data were obtained and determined that there is no evidence for time-dependent behavior within the data and there is no evidence that the data do not follow a Poisson distribution. The optimal count time varies from sample to sample, as it was determined to be a function of the desired level of analytical uncertainty. Sr results for a speleothem obtained through LA-ICP-OES show a high degree of comparability to results obtained through µ-XRF and demonstrate the suitability of this non-destructive method. Sr counts can also be converted to concentration [ppm] via Xray-tube-dependent calibration curves.

Rapid measurement of strontium in speleothems using core-scanning micro X-ray fluorescence

“…Where sampling is retrospective, such as in this study, the opposite face of the sampled stalagmite can be used, with care taken over the additional chronological uncertainty. Buckles and Rowe (2016) suggest that samples should be polished, to minimize the impacts of surface roughness on elemental bias, as the characteristic x-rays of lighter elements are emitted from shallow depths. This "information depth", defined as where 63% of the secondary x-rays escape the matrix, is dependent on several variables including the density, porosity and fabric of the carbonate matrix and the sample geometry.…”

“…2). The distribution of Sr, Ca and Sr/Ca are expected to form a Poisson distribution (Buckles and Rowe, 2016), but are also statistically indistinguishable from a normal distribution. Sr counts had a 2-standard deviation range (2) of 354 counts or 4.1% of the total.…”

“…Li et al (2015) show an order of magnitude offset between Itrax Sr/Ca and ICP-OES Sr/Ca. However, Buckles and Rowe (2016), show little offset between their portable-XRF system and ICP-OES. This suggests that the Itrax offset is not inherent to XRF analysis but rather arises from incorrect calibration in the signal processing specific to this methodology.…”

“…Non-vacuum XRF allows for larger samples, but the attenuation of secondary x-rays in the air-gap prevents the analysis of lighter elements such as Mg, which attenuates 90% in 1cm of air. Buckles and Rowe (2016) used a portable XRF system on speleothems, showing non-vacuum XRF to be a viable option for analyzing trace elements in speleothems. In our study, we build on the work of Buckles and Rowe (2016), providing additional internal tests, external comparisons with other methodological techniques, and using a different machine, with greater lateral scan length.…”

The karst and its neighbors: Digital map Multi-year cave dripwater frequency and hydrochemical monitoring of three caves in Eastern North America

“…Culverson Creek Cave (CCC), Buckeye Creek Cave (BCC) and Lost World Caverns (LWC), in Greenbrier County, southeastern West Virginia (Fig. 1), were selected for a monitoring study for several reasons: 1) stalagmites from CCC and BCC have been used in numerous paleoclimate studies (Springer et al, 2008(Springer et al, , 2009(Springer et al, , 2014Hardt et al, 2010;Buckles and Rowe, 2016), 2) the caves are located near one another with relative ease of access, and 3) the caves exhibit a wide range of morphologies and drip behaviors. These caves are formed within the Union Limestone Member (massive limestone with abundant chert and marine fossils) of the Greenbrier Group, which is Mississippian in age and is composed of interbedded limestone, shale, and sandstone (Dasher and Balfour, 1994).…”