The project has fully achieved its objectives and milestones. The following work was carried out in order to achieve the three specific objectives of GOLD-ICE:
1. The first specific objective (SO1) was to characterize a LA-ICP-MS setup at UNIVE specifically dedicated to ice core analysis
GOLD-ICE has designed a cryogenic sample holder specifically dedicated to holding strips of ice cores, artificial ice as well as glass standards.
Special focus was the integration of a “aerosol rapid introduction system” – allowing to reduce the washout time from seconds to milli-second, i.e. the time needed for transferring the ablated sample aerosol plume to the ICP-MS. With washout times in the millisecond range, the recording of baseline-separated single pulses at high repetition rates becomes possible – the GOLD-ICE method uses almost 300 shots per second (previous ice core LA-ICP-MS studies: 10 – 20 shots per second).
As a result, the scan speed employed in the GOLD-ICE method is roughly 20 times higher with respect to previous studies on ice cores – now reaching around 1 millimeter per second. This is the basis to employ state-of-the-art 2D LA-ICP-MS imaging, allowing to move much faster with the laser, and to generate artifact-free images in realistic time with higher spatial resolution: 35µm.
Various elements have been tested for the new imaging application. Due to constraints by the imaging method, four elements are measured at a time. Usually these elements are Na, Mg, Mn and Sr – all of interest with respect to ice core applications.
2. The second specific objective (SO2) was to explore the potential and limitations to a paleo-climatic interpretation of LA-ICP-MS data in view of the ice core micro stratigraphy
With the integrated camera mosaic of visual images of the ice sample surface were obtained. The comparison of the impurity images generated by LA-ICP-MS with these mosaics allowed to assess clearly the localization of impurities in the ice crystal matrix. GOLD-ICE revealed an unequivocal correlation between high impurity intensities and the locations of grain boundaries, most pronounced for Na. This affinity of Na to be located at grain boundaries is consistent with previous investigations with other, non-imaging methods showing how LA-ICP-MS can provide a complementary tool for investigating the localization of impurities in glacier ice.
Consequences of the impurity localization with respect to the interpretation of individual LAICP-MS lines measured along the main core axis were assessed: Due to the localization at boundaries, the grain network strongly affects the high-frequency variability of the LA-ICP-MS signal: Individual peaks occur when the laser beam intersects a grain boundary. Accordingly, GOLD-ICE demonstrated that, in such cases, the significance of single peaks (the high-frequency signal components) is tied closely to the ice crystal network, rather than to individual past climate-related events.
3. The third specific objective (SO3) was to fully exploit the high resolution of LA-ICP-MS for the analysis of the oldest layers in polar ice cores
The new imaging approach of GOLD-ICE delivered an improved understanding of the LA-ICP-MS signal origin in ice cores. GOLD-ICE showed that the influence of the grain boundary network depends on the type of impurity (e.g. strong for Na, weaker for Mg) and on spatial resolution. At coarser spatial resolution the imprint of individual crystal features weakens – making the choice of resolution and scale a cornerstone of the experimental design in LA-ICP-MS ice core analysis.
Several large images of deep ice in Antarctic ice cores were obtained (EPICA Dome C and Talos Dome). Samples were chosen to represent chemical and physical properties characteristic for deep ice and both glacial and interglacial periods, e.g. with comparatively large and small grain sizes, respectively.
Advanced statistical methods have been employed for data analysis. It is shown how an experimental design can be adapted targeting to disentangle climate- and crystal-related signal components in the LA-ICP-MS analysis.
