Reproduction of melting behavior for vitrified hillforts based on amphibolite, granite, and basalt lithologies

European Bronze and Iron Age vitrified hillforts have been known since the 1700s, but archaeological interpretations regarding their function and use are still debated. We carried out a series of experiments to constrain conditions that led to the vitrification of the inner wall rocks in the hillfort at Broborg, Sweden. Potential source rocks were collected locally and heat treated in the laboratory, varying maximum temperature, cooling rate, and starting particle size. Crystalline and amorphous phases were quantified using X-ray diffraction both in situ, during heating and cooling, and ex situ, after heating and quenching. Textures, phases, and glass compositions obtained were compared with those for rock samples from the vitrified part of the wall, as well as with equilibrium crystallization calculations. ‘Dark glass’ and its associated minerals formed from amphibolite or dolerite rocks melted at 1000–1200 °C under reducing atmosphere then slow cooled. ‘Clear glass’ formed from non-equilibrium partial melting of feldspar in granitoid rocks. This study aids archaeological forensic investigation of vitrified hillforts and interpretation of source rock material by mapping mineralogical changes and glass production under various heating conditions.

S-4 S1.2 Electron probe microanalysis (EPMA) linescan of Bro-W JEOL JXA-8500F, WSU GeoAnalytical Lab. ZAF (Armstrong-Love/Scott) intensity correction. 15 kV, 10 nA, 10 μm beam size. *Acquired using Thermo UltraDry EDS spectrometer. Peak counting time is dead-time corrected. S2.2 Thin sections Petrographic analyses of archaeological samples and source rocks were performed on thin sections using an Olympus BX53-P (archaeological samples, dike, and granites) or Zeiss Axioskop 40A (amphibolites) polarizing light microscopes equipped with integrated digital cameras for continuous documentation of the analyses. Thin sections were prepared by Spectrum Petrographics (Vancouver, WA, USA) or Axinit (Bratislava, Slovakia). . WG shows evidence of being previously heated and recrystallized. Granitoids "442" and "443" (see Table S-3) were also investigated. 442 was similar to WG and 443 was similar to RG. S-9 S3 Ex situ XRD and Rietveld refinement

S3.1 Methodology
Ex situ XRD. X-ray diffraction of ex situ archaeological samples, source rocks, and experimentally melted rocks was performed using a PANalytical MPD PRO X-ray diffractometer outfitted with a Co Kα X-ray tube with 40 kV accelerating voltage and 40 mA and an X'celerator detector. The scan parameters were 2-90° (Co) measurement range, 0.05° step size, 10 s dwell, and 15 repetitions per sample.
Rietveld refinement was performed using HighScore plus (PANalytical, Netherlands) on powdered materials either without an internal standard but with a predetermined background or with a 20 wt.% corundum standard. Refinements included scale factor, specimen displacement, unit cell, u/v/w parameters, peak shape, split width and shape, and preferred orientation (using spherical harmonics).
Some samples were also measured with a Cu Kα X-ray tube on the same instrument, and the resulting measurements and refinements were comparable.
Powders were obtained for source rocks by crushing to smaller sizes to obtain a uniform representation of the heterogeneous rocks. Note that this allows determination of average mineral composition of the rocks (this was particularly observed for BA3 which had large quartz seams). This should not be confused with melting experiments on powders; melting of large rocks of these large grain-size materials (> tens of micrometers) would necessarily be different than melting of powders.

S3.2 Amphibolites
The monolith samples were placed in Pt containers and placed into a furnace at 850°C or 1050°C, and left for 15 min before air quenching. Air quenching of monolithic samples, at 850 and 1050°C generally caused them to crumble, and this was determined to be due to thermal shock of the monoliths rather than from any reaction of the phases. Rietveld refinements for BA5 (site 5) amphibolite with Co tube data. Room temperature assumed as 20°C. All samples except room temperature rock used an internal standard.
Table S-4. XRD analysis results (wt.%) for BA5 amphibolite at 1200°C with different crucible and cooling conditions. Notes (see Methods for details): 'slow' cooling is 10°C min -1 ; 'chunks' are 2-3 cm monoliths; 'chips' are <3 mm particles.    S-16 S3.4 Basalts Starting material for UMAT1 was powdered rock while for BGR it was chips. All were put in at temperature, held for 15 minutes, then removed from the furnace to air quench. Room temperature is assumed to be 20°C.

S4.1 Methodology
In situ hot stage XRD of source rocks. In situ hot-stage XRD of powdered samples was carried out using a Bruker D8 Advance X-ray diffractometer (XRD) outfitted with a Cu Kα X-ray tube (tube parameters: 45 kV and 40 mA) and a platinum heating stage apparatus. The scan parameters were 2-70° 2θ measurement range, 0.01° step size, 20 s dwell. Variable divergence slits were used to restrict the beam width to 10 mm, due to uneven heating beyond the center of the platinum heating strip caused by preexisting thermomechanical creep, which led to necking of the heating strip (Fig. S-19). Samples were first crushed using an agate mortar and pestle then the powder was deposited onto the heating strip using ethanol and a plastic pipette. In situ diffraction measurements were performed on selected samples (BA5 amphibolite, dike dolerite, HAS37 basalt) both while heating from room temperature and while cooling from elevated temperature. In situ heating trials were carried out by heating samples from room temperature to 1200°C at a heating rate of 10°C min -1 with 1 h dwells at 200, 500, 700, 850, 950, 1050, and 1200°C to allow for XRD measurements. In situ cooling trials were carried out by heating samples from room temperature to 1500°C at a heating rate of 10°C s -1 then cooling from 1500°C to room temperature at 10 C min -1 with 1 h dwells at 1450°C and 50°C increments from 1250-800°C to allow for XRD measurements. Fig. S-20 shows the cooling and heating profile used for this study.

Fig. S-20.
Cooling and heating profile for hot-stage XRD Rietveld refinement and peak profile fitting was carried out using HighScore Plus (PANalytical) software. Any addition of an internal standard would likely interact and alter the chemistry throughout the duration of the high temperature measurement; therefore, another method had to be developed to estimate the amorphous fraction as a function of temperature (see Results). The method chosen first involved obtaining the peak area of one phase (anorthite) as a function of temperature which was then fit to the Johnson-Mehl-Avrami-Kolmogorov (JMAK) equation:

Equation (1)
where A/A 0 is the peak area of the anorthite phase normalized to the average-measured using a Calgliotti profile fitting function)-T is the dependent-variable temperature in degrees Kelvin), τ is the temperature at the inflection point (in degrees Kelvin), and n is the so-called Avrami exponent. The value of n was 17 for all samples and the τ value was 1480 K for BA5 amphibolite, 1368 K for the dike dolerite sample, and 1391 K for the HAS37 basalt. After fitting the JMAK equation, the resulting functions were used to compute a temperature-specific correction factor, F T , such that F T =C 0 (A/A 0 )/C T , where C 0 is the average fraction of anorthite prior to melting and C T is the fraction of anorthite at the specified temperature (i.e., 1050°C or 1200°C). The correction factor was multiplied to all phases and the amorphous fraction in the sample was found by the difference method.
This correction was modified for the case of the dike sample at 1200°C, where no anorthite was detected and the only phase present was magnetite. As a result, applying the correction factor would result in infinite magnetite. Therefore, the correction factor was instead taken as the value of A/A 0 from the JMAK fit for the temperature of 1200°C, which provided a reasonable estimate of the amorphous fraction at 1200°C for the dike sample. S-21 S4.2 In situ vs. ex situ data, extended discussion S4.2.1 Amphibolite (BA5) Amphibole (hornblende) and feldspar have similar trends in both measurements. The amphibole fraction shows a sharp decrease at 950-1050°C (in situ), comparable to the large drop at 850-1050°C (ex situ), with phase fraction at zero at 1200°C in both measurements. Similarly, feldspar fraction drops at 1000°C in in situ but increases 1050°C and decreases 1200°C in ex situ. This difference is due to the requirement of fitting the feldspar to obtain the glassy (melt) fraction in the in situ run, as well as the fact that new feldspar forms on cooling in the ex situ experiment. To assess the minimum temperature for glass formation in the ex situ experiment, a 750°C data point was collected, showing no glass formation; thus melting begins between 750°C and 850°C. For the in situ experiment, better resolution can be seen regarding the breakdown of phyllosilicates (biotite, chlorite), which is absent between 500°C and 700°Ccompare to Fig. 4 where some biotite remains after 15 min at 850°C. Also, quartz undergoes its low to high transition between 500°C and 700°C, which is not seen in the ex situ data since the quartz observed is always at room temperature (i.e., low quartz). In the ex situ experiment, some quartz still persists even after treatment at the highest temperatures, which is either related to kinetics (longer dwell times for in situ) or, less likely, due to new quartz being formed upon cooling.
A small amount of spinel (magnetite) is seen at the highest temperatures in the in situ experiment, with much more seen with ex situ, which is consistent with the finding above that most magnetite forms upon cooling. Similarly, no pyroxene is observed in situ, but is observed ex situ, since it forms on cooling. This result was again confirmed from the in situ XRD cooling measurements, where magnetite formed first upon cooling to 1100°C, with pyroxene and hematite forming at 1000°C. In the MELTS simulations (Fig S-22), which are equilibrium calculations, magnetite forms below 1400°C at the hematite-magnetite (Hm-Mt) buffer, but only below 1200°C for the fayalite-magnetite-quartz (FMQ) buffer. Hematite forms below 1100°C in Hm-Mt buffer, but only in trace amounts in FMQ buffer below 900°C. Pyroxene forms below 1200°C in both redox conditions, but in larger amounts in FMQ. Olivine only forms in FMQ conditions, below 1200°C, confirming the need for reducing conditions for olivine, as found in the ex situ experiments. S4.2.2 Dolerite (dike) In the case of the dolerite, the match between the amphibole and feldspar temperature trends is more similar for the two measurements. The formation of the melt is sharp between 950°C and 1050°C in situ, while some glass forms even at 850°C in the ex situ experiment, again similar to the amphibolite results; a 750°C data point showed no glass formation, thus melting begins between 750°C and 850°C. In the in situ results, the minor clinochlore phase is gone by about 550°C, and the minor quartz above 1050°C. High temperature phases include very minor spinel in the in situ case at 1200°C, whereas all feldspar and amphibolite have been destroyed. This was similar in the ex situ case, but spinel forms for the 1050°C max heat-likely upon cooling-but at 1200°C max heat the preferred iron oxide phase is hematite. Pyroxene only forms in the ex situ case, on cooling, with more forming when maximum temperature is 1050°C than when it is 1200°C.
In the in situ cooling experiments, both magnetite and pyroxene formed below 1100°C, with pyroxene being the more dominant phase-as determined by signal intensity-until the temperature dropped to 800°C, when magnetite became the dominant phase. In the MELTS simulations (Fig S-23), magnetite was the liquidus phase below 1400°C for Hm-Mt buffer, but olivine was the liquidus phase below 1350°C for FMQ buffer. Hematite was only predicted to form in the Hm-Mt buffer, below 1000°C. Thus, apparently the experiment was slightly oxidizing for the ex situ samples than for the in situ ones. In both redox conditions, feldspar and pyroxene crystallize below 1200°C. It should be noted that in no case was feldspar observed in the in situ cooling experiments of melted rocks, despite its prediction from equilibrium thermodynamics. This is due to the well-known problem of the difficulty of homogeneous nucleation of feldspar. S4 S-22           Table 1 in the main text.