RECONSTRUCTING VITRIFIED WALLBUILDING TECHNOLOGY FOR BROBORG HILLFORT SITE, SWEDEN

One primary goal of this international collaboration is to utilize the natural experiment of long term glass-ceramic degradation at the Broborg Hillfort in Sweden to inform models incorporating compositional dependence of glass corrosion. In tandem with this goal, we are supporting the work to preserve the cultural heritage and rediscover the technology of the pre-Viking group whose practices lead to the unusual natural experiments. To this end, we have investigated the micro-scale interfaces between different glass and crystalline phases in some samples from Broborg using X-ray diffraction and Electron Probe Microanalysis.

Microstructures observed indicate high undercooling resulting in dendritic textures of pyroxenes, feldspars, and cordierite. Relict and newly formed spinels containing iron have also been considered as recorders of the Earth’s magnetic field, and preliminary experiments using a Vibrating Sample Magnetometer have determined that glassy hillfort specimens are suitable for paleomagnetic measurements, including irrefutable tests of whether melting was due to lightning or anthropogenic heating.

Experiments conducted to replicate the glassy phase, for the purposes of conducting corrosion experiments, indicated that excessively high temperatures >1450°C were required to melt compositions measured in hillfort samples as both “dark” (mafic) and “light” (felsic) glasses. These temperatures would not be achievable in antiquity, so further investigations have begun to assess the role of water content and iron redox in melt viscosity. Additionally, samples of the host rock amphibolite, thought to be the source of the mafic melt, have been obtained from Sweden and characterized for chemistry and phase. Initial investigations suggest than not all local amphibolites would have been suitable, with some having excessively high quartz content and inadequate amphibole minerals containing bound hydroxides to lower melting temperatures. Characterization of hillfort samples, simulant glasses, and amphibolite rocks will be presented. Production of simulant glasses with the right character and melting point is critical for accelerated lab corrosion testing to compare with the natural analogue glasses, exposed for ~1500 years.