Using FTIR-ATR, analytical colour and mercury for unravelling the cremation ritual of Tyresta Viking Age burial mound (South-Central Sweden)

doi:10.21203/rs.3.rs-7273566/v1

Download PDF

Research Article

Using FTIR-ATR, analytical colour and mercury for unravelling the cremation ritual of Tyresta Viking Age burial mound (South-Central Sweden)

https://doi.org/10.21203/rs.3.rs-7273566/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

The funerary rituals in Viking Age Scandinavia are known by their complexity and diversity including inhumation, boat burials, and cremation. Cremations have been extensively studied macroscopically, but the application of geochemical techniques, although highly informative to investigate cremation in more detail, had not been widely explored yet. In the Viking Age the inclusion of both animals and humans on the funeral pyres makes its research challenging. In the present study, we demonstrate the usefulness of molecular analysis (FTIR-ATR), direct mercury analysis (DMA) and analytical colour (in the CIELab space) for the characterization of Viking Age cremated remains of a human individual buried with dogs, a horse, a cat and other animals – as well as evaluate naked-eye methods for assessing the temperature of firing. We demonstrate that the spectroscopical signal is highly correlated with analytical colour parameters, and with mercury concentration (that still present even in bones exposed at high temperatures). Mercury concentration was higher in human bones than animals, suggesting an ante-mortem pollution. Human bones were probably heated at temperatures lower than 1100°C but higher than 900°C, while animals were cremated at lower temperature. We also identified a secondary cremation in a layer previously interpreted as a Badger burrow. A new MIR-index, _TPV, is proposed, and we suggest that the cyanamide band is related to oxygen availability during burning. Regarding naked-eye colour method, we support the splitting into two categories the white colour degree regarding the porosity of the bone, to better approach higher temperatures.

cremated remains

Viking age

FTIR-ATR

mercury

Analytical colour

The cremation ritual has been widely used in the past. Burning a human body was the most common form of funerary ritual in many Norse communities, during the Viking period (Price and Raffield 2024). In the Nordic Late Iron Age (AD 800–1050), mortuary practices were characterized by diversity, combining inhumation and cremations, with the latter occurring more frequently until the arrival of Christianity (Price and Raffield 2024). During ritual, the deceased, often accompanied by grave goods and animal offerings, was placed on a pyre to be burned; more grave goods could be added once burned, followed by the remains being manipulated. The burnt skeletal remains could also be relocated into a pit, a container or spread out and in many cases a mound made the grave visible in the landscape. The number of animals and species included on funeral pyres increases during this period compared to the Early Iron Age, and socially prestigious animals, such as horses, were common (Holck 1997; Iregren 1997; Sigvallius 1994; Bratt 2008). The performance of the cremation and the level of firing of the skeletal remains, as judged from macroscopic observations, may vary between species, and from one burial to the other. In addition, several graves display post-depositional manipulations that suggest contemporary or later intrusions. All these facts increase the complexity of excavating and studying cremation burials from Norse people but also challenge researchers to search for new tools to understand Viking period rituality.

Fire, even at low temperature, burns soft tissues and often severely damages the skeleton, causing fractures and transformations. Recording and collecting cremated remains is not easy, and the study of cremations was traditionally seen as secondary of peripheral (Thompson 2015; Williams 2008). The fragmentary nature of cremated remains, or their poor collection on the field, is compounded by the challenges of adapting laboratory methodologies to their study (Thompson 2015). Also, new archaeological questions related with the combustion process have arisen, such as the skills of those performing the cremation (e.g., Kjellström et al. 2024). This aspect of the performance of the cremation would probably be reflected by the level of firing of the skeletal remains. Temperature-induced colour changes in bone had been extensively studied in the literature as a tool for assessing factors such as the temperature achieved, time of heating, firing pattern (see Holck 1997, Stiner et al. 1995, Symes et al. 2015; Gallo et al. 2023). It has been demonstrated that there is a relation between colour changes and chemical-structural changes in bone with burning (Symes et al. 2015; Gallo et al. 2023). However, most classification methods are based on visual/macroscopic features, which often are highly subjective (Thompson et al. 2017). The presence of manganese and iron oxides, clothing, objects, or oxygen availability can also affect the colour of burnt bone (Shahack-Gross et al. 1997; Michel et a. 2006; Innocenti et al. 2024). Therefore, there is a need in developing more objective methods for the study of cremations. In recent years, (bio)geochemical techniques for the study of archaeological skeletal remains, such as FTIR-ATR, have been employed for comparing trends in the variation of the spectral signal considering different variables such as bone type, temperature, sex, age, species, fuel used, burning environment. However, much remains to be done, as current studies are often based on a few samples per individual, have little statistical robustness, and researchers relate observed changes directly to the effect of fire, without considering the ante-mortem signal.

In this paper, we analyse the Tyresta cairn, located south of Stockholm (Sweden), where a large number of animal remains were found along with what is believed to be a single cremation of a human being. Also, the burial mound was erected at the place of the funeral pyre. We aim to: i) test how accurate macroscopical colour classification is by comparing it with analytical colour analyses, and geochemical analysis (mid-infrared (MIR) Fourier transform, attenuated total reflectance, spectroscopy FTIR-ATR), ii) characterise Tyresta cremation by assessing structural-chemical changes by FTIR-ATR (MIR-indices and single bands in multivariate analyses), and iii) explore a new approach to temperature estimation in archaeological cremations such as total mercury content, since this element is highly volatile and temperature-sensitive Indeed, to our knowledge, the potential of cremated remains for the study of mercury paleo-pollution is addressed here for the first time. Both ante- and post-mortem characteristics will be considered following the moto of our previous investigation regarding FTIR in inhumated bone (Colmenares-Prado et al. 2025). Our main objective is to deepen into the pre-Christian Norse world and the relationship between the person and the animals that shared a grave on their final journey.

2.1. Tyresta Archaeological site

The cremation studied here was part of a Late Iron Age burial mound excavated in 2012 and located in Tyresta village (~ 20 km south of Stockholm, Sweden) (Fig. 1) A radiocarbon (C14) dating of a grain yielded the range 1129 ± 31 BP, i.e., the Viking Age, which calibrated most likely falls between 780–990 CE (95,4% - Ua-44723). The mound (~ 12 m in diameter and 2 m in height) was constructed on top of the funeral pyre and the cremation remains. A central cairn (~ 3 m in diameter) was constructed on top of the funeral pyre and was covered by a sand mantle with a clay embankment. In the filling of the mound there were two secondary features with large stones and secondary deposits of cremated remains (Sec. Burial 1 and 2), as well as a third larger secondary deposit of cremation remains (labelled as Badger layer) that had been placed on the original surface of the mound with some tunnels extending into it, near the central cairn. A cellar from 18-19th centuries CE had been built into the mound disturbing a portion of the outer parts, but it did not reach the central cairn (Fig. 2). There were no traces of earlier Iron Age activities or older graves beneath the burial mound and the funeral pyre.

Large pieces of charred wood (blanks and logs used as fuel) among the funerary remains suggest that the mound was erected on top of the remains of the pyre. Thus, initially, a rectangular (6x5 m) pyre had been constructed. After the human and animal bodies and grave goods were placed and the construction finalized, the pyre was burnt down. After the cremation a few large human bone fragments were first collected and placed in two burial vessels that were deposited at the bottom where the central cairn was built. The cremation layer was thickest beneath (and in) the central cairn and thinner in the peripheral parts beneath the mound, suggesting that the cremation remains had been swept towards the central part where the central cairn was constructed on top of the remains. In the excavations, the cairn was excavated in four technical layers (three cremation layers (S1-S3) and a layer (ÖB) immediately above S1 including the filling in the top of the cairn, in which some burnt bones were also present (Fig. 2).

Apart from the cremation remains, artifacts and ecofacts were found: six glass beads, a spindle whorl, ceramic vessels (and finds of sintered pottery), gaming dice and pieces, a comb, a fragment of a lathed wooden vessel, decorated wooden objects, tools with wooden handles (probably from knives), charred remains of textiles and 74 iron rivets - that suggests that a boat or parts of a boat may have been on the pyre)- In addition, ‘food’ was also found that included bread, fish, and grains (hulled barley).

A total of c. 6370 g of skeletal remains was recovered, and 4296 fragments were identified and classified by species or group of species, Supplementary Table 1. The analysis followed standard methodologies developed and adapted for cremated remains (see e.g., McKinley 2015, or Sigvallius 1994). The methodology is described in detail in the excavation report (Pettersson, personal communication). The study of skeletal remains allowed the identification of a human individual (probably a female adult), a horse, three dogs, a cat, a sheep/goat, a goose, a galliform bird, and one fish bone. In addition to these species, some fragments were classified as large, medium or small sized mammal. There were important differences in level of firing between the species. Unburned remains of two dogs were identified in the central cairn and a third one (cremated) was found in the funerary. Finally, the cremated remains of a fourth small sized dog were found in the Badger layer. Levels of firing up to level 3 and 4 were identified among bones of human, horse, dog, cat, and bird (Pettersson, personal communication).

2.2. Material

A total of 107 cortical bone samples from Tyresta archaeological site were analysed by macroscopical colour and FTIR-ATR (non-destructive); of those, 21 were selected for analytical colour and mercury (destructive/invasive analysis. Bone fragments macroscopically identified as belonging to four different species (human n = 40, horse n = 26, dog n = 38, and cat n = 3) were selected. Additionally, information regarding bone type was considered classifying them into four types (bones from hands/feet n = 23, crania including jaw n = 29, limb bones including long bones and one patella n = 31, and thoracic bones including vertebrae and ribs n = 23). Gonçalves et al. (2018) recommend using the same bone type, age, and bone region when sampling for a comparative study. However, in cremation burials, the preservation of the bones is not homogeneous, and the bones are not as easily identifiable. Therefore, and for obtaining a statistically significant number of samples, the bones for analysis were selected based on availability considering different degrees of burning, layers of the cairn, and species, in order to address as most variability of the pyre as possible. Five layers were investigated in the pyre: four inside the cairn (ÖB > S1 > S2 > S3), and one outside it (Badger layer). The distribution of the layers and bone species in the different layers is shown in Fig. 2.

2.3. Macroscopic Colour

A macroscopic classification was performed following a modification of Stiner et al. (1995) (m-Stiner). These authors established 7 degrees of burning (0–6) that represent a gradient of transformations in which 0 is completely unburnt; 3 is carbonized but not calcined (black and brittle); and 6 is fully calcined (white and crystalline). The other categories represent intermediate degrees. The method was modified according to the Osteoarchaeological Research Laboratory of Stockholm University standards. This was done by adding an extra category inspired on Hermann’s (1988) observations, i.e., splitting the fully calcined bone category (Stiner’s grade 6) in two: fully calcined with chalky appearance -soft and powdery- (grade 6A) and fully calcined with porcelain-like appearance -crystalline appearance and sound-(grade 6B).

2.4. Analytical colour

A total of 21 cortical bone samples were cleaned in an ultrasound bath (ultrapure water type 1, low metal trace, at least 5 times each) and finely milled (~ 20 mg) using a Retsch mixer mill MM301. Colour was measured in the CIElab colour space with a Konica-Minolta CR-5 Chroma Meter for solids (Konica Minolta Inc., Japan). CIELab space is a three-dimensional space in which the following parameters are measured: L* (luminosity: from 0 to 100; black to white), a* (gradient from green (negative values) to red (positive values)), b* (gradient from blue (negative values) to yellow). Also, another three-dimensional space was considered: L (luminosity), C_ab* (chroma: it increases from the centre of the sphere to the exterior), and h_ab (hue; angle between a* and b*).

2.5. Mercury concentration

Total bulk mercury concentrations in 21 samples from human and animal bone were determined in fine-milled bone using a DMA-80 Milestone triple-cell Hg analyser hosted at Ecopast low-metal-trace-lab at CRETUS Institute Universidade Santiago de Compostela). We followed the protocol described in (Álvarez-Fernández et al. 2020; López-Costas et al. 2020). Average analysed weight per sample is 17.9 ± 4.4 mg of milled bone. Replicates were made in every sample obtaining good replicability (coefficient of variation lower than 10%, 3.4 ± 3.6%, and/or low raw difference between both replicas below 1 ng g^− 1, 0.3 ± 0.3 ng g^− 1). Two standard reference materials were measured alongside the samples: bone ash NIST1400 (0.92 ± 0.08 µg kg^− 1, 4 times) and bone meal NIST 1486 (3.36 ± 0.47 ng g^− 1, 4 times), the last reached 100% recovery. A total of 30 blanks were run with the samples obtaining in all cases values below detection limit (0.07 ± 0.01 ng g^− 1). Several reference materials as well as internal standards are analysed weekly, all reaching 100% recovery. The quantification limit is 0.5 ng g^− 1. No water was added to the samples, since bone is a non-inflammable material.

2.6 FTIR-ATR

Cortical bone samples (~ 1–3 mg) were obtained by drilling the selected bones with a dental equipment. The samples were analysed by FTIR-ATR using a Termo Scientific Nicolet iS10 FTIR equipped with a diamond crystal ATR accessory hosted at the Archaeological Research Laboratory at Stockholm University. The IR-spectra were recorded in the range between 4000 and 525 cm^− 1, 100 scans per sample were recorded with a resolution of 4 cm^− 1, and a background was collected before every measurement. Spectra were smoothed with the OMNIC 8.2.387 software. Bone powder was recovered from the equipment after every measurement and stored in Eppendorf tubes.

Spectra were SNV normalized, and second derivative spectra were obtained using Orange v3.38 data mining software (Demšar et al. 2013). No baseline correction was performed, since this procedure lowers the 600 − 525 cm^− 1 region, affecting the relative peak intensities, and therefore affects the MIR-indices calculations (See supplementary Fig. 1).

MIR-indices based on literature were calculated using band intensities (Table 1):

Table 1
MIR-indices calculated here with its formula, meaning and references
MIR-index	Formula	Meaning	Reference	Observations
Infrared splitting factor (IRSF)	\(\:\frac{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim560\:{\text{c}\text{m}}^{-1}\:+\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim600\:{\text{c}\text{m}}^{-1}}{\text{m}\text{i}\text{n}\text{A}\text{b}\text{s}\:\sim595\:{\text{c}\text{m}}^{-1}}\)	It is an indicator of the crystallinity of bioapatite. The higher the index, the more crystallized the bioapatite is.	Weiner and Bar-Yosef 1990	Care must be taken when interpreting these results, since crystallinity (and this index) not only increases with burning, but also with diagenesis, Stiner et al., 1995).
Carbonate-to-phosphate (C/P)	\(\:\frac{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim1415\:{\text{c}\text{m}}^{-1}}{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim1015\:{\text{c}\text{m}}^{-1}}\)	Carbonate content relative to phosphate	Grunenwald et al. 2014	We use here the ν₃ CO₃²⁻ and ν₃PO₄³⁻domains (Grunenwald et al., 2014). Care must be taken when interpreting these results, since carbonate loss (and this index) not only occurs with burning, but also with diagenesis, Stiner et al., 1995).
B-type carbonate-to-phosphate (BPI)	\(\:\frac{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim1415\:{\text{c}\text{m}}^{-1}}{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim600\:{\text{c}\text{m}}^{-1}}\)	Carbonate type B content relative to ν₄ phosphate	LeGeros and LeGeros 1983 in Snoeck et al. 2014a
A-type-carbonate-to-phosphate (API)	\(\:\frac{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim1540\:{\text{c}\text{m}}^{-1}}{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim600\:{\text{c}\text{m}}^{-1}}\)		Sponheimer and Lee-Thorp 1999
A-type to B-type carbonate (C/C)	\(\:\frac{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim1445\:{\text{c}\text{m}}^{-1}}{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim1415\:{\text{c}\text{m}}^{-1}}\)	Carbonate type A relative to carbonate type B content	Snoeck et al. 2014a
Amide-to-phosphate (Am/P)	\(\:\frac{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim1660\:{\text{c}\text{m}}^{-1}}{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim1015\:{\text{c}\text{m}}^{-1}}\)	This index is an indicator of the collagen relative to phosphate in the bone	Trueman et al. 2004
Protein-to-carbonate (CO/CO₃)	\(\:\frac{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim1650\:{\text{c}\text{m}}^{-1}}{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim1415\:{\text{c}\text{m}}^{-1}}\)	Protein content relative to carbonate content	Thompson et al. 2013
OH⁻ relative to phosphate (OH/P)	\(\:\frac{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim630\:{\text{c}\text{m}}^{-1}}{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim600\:{\text{c}\text{m}}^{-1}}\)	This index is an indicator of changes in the content of hydroxyl groups.	Snoeck et al. 2014a
Cyanamide-to-phosphate (CN/P)	\(\:\frac{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim2010\:{\text{c}\text{m}}^{-1}}{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim1015\:{\text{c}\text{m}}^{-1}}\)		Zazzo et al. 2013
Phosphate high temperature (PHT)	\(\:\frac{\text{m}\text{a}\text{x}\text{A}\text{b}\text{s}\:\sim625\:{\text{c}\text{m}}^{-1}}{\text{m}\text{i}\text{n}\text{A}\text{b}\text{s}\:\sim610\:{\text{c}\text{m}}^{-1}}\)		Thompson et al. 2013
Full width at half maximum of the ν₃PO₄³⁻ peak (FWHM)			Thompson et al. 2013
Temperature-phosphate-valley (_TPV)	\(\:\frac{maxAbs\:\sim1088\:c{m}^{-1}}{\text{min}Abs\:\sim1077\:c{m}^{-1}}\)	Sharpening of the ν₃ phosphate at ~ 1088 cm^− 1: absorbance ratio at ~ 1088 cm^− 1 and its valley respect to the main ν₃ phosphate band	This study

2.6. Statistical analysis

A principal component analysis was performed on selected peaks of the normalized spectra. Peak selection was made by RStudio {Andurinha} package (Álvarez Fernández and Martínez Cortizas 2020) using the standardized spectra and the second derivative spectra. As described below, both the values of the MIR indices and a preliminary PCA, showed that the whole bone sample was composed of two separate groups, which were confirmed by discriminant analysis. Thereby, two additional PCA were performed to the two different populations separately.

Samples were statistically analysed by the following categories: species (human, horse, dog, cat), bone type (cranium/mandible, long bone, hand/foot, thoracic bone), and the m-Stiner cremation degree (0-6B). Normality was tested on the samples using Shapiro-Wilks and since distributions did not follow a normal distribution, non-parametric Kruskal-Wallis test was performed to check for statistically significance among categories between the indices and PCs. Tests with significance p < 0.05 were considered.

3.1. Macroscopic colour

The majority of the 107 samples correspond to m-Stiner grade 6A (n = 37) and 6B (n = 33). However, there are some samples belonging to grade 0 (n = 13), grade 2 (n = 2), grade 3 (n = 4), grade 4 (n = 7), and grade 5 (n = 11). In Fig. 3 the distribution of samples for species and m-Stiner degree is shown.

3.1.2. Analytical colour

Average and range CIELab values for the measured sample selection (n = 21) (separated in unburnt samples (m-Stiner grade 0) and burnt samples m-Stiner grade > 0) are displayed in Table 2. All the colour parameters’ values are shown in Supplementary Table 2. Note that redness (positive a* values) is higher for samples from layer S3 than layer S2 (p = 0.005); yellowness (positive b* values) is higher for samples from layer S3 than those from S2 (p = 0.004); and chromaticity (C*) is higher in samples from S3 than those from S2 (p = 0.004). No differences in colour were shown by bone type. Human bones present higher lightness (L*), and hue (h°) but lower red component (a*) compared with dog bones. These results are mainly due to the fact that most dog bones are classified as unburnt (i.e., non-cremated) while human bones are mainly classified as calcined (i.e., intensely cremated). Parameters a* and b* according to species are represented in. 4 and Fig. 5. L* is higher for samples with m-Stiner grade 6A compared with grade 0, meaning that luminosity is higher in calcined samples than in unburnt samples. Also, m-Stiner grade 0 presents higher a*, b* and C* values compared with m-Stiner grade 6B.

Table 2
CIELab average and range parameters for unburnt (m-Stiner = 0) and burnt samples. (m-Stiner grade > 0)
Unburnt	L*	a*	b*	C*	h
Range	59.49–64.39	4.92–5.66	16.43–19.26	17.2-19.97	72.78–74.67
AVG	62.35 ± 1.83	5.24 ± 0.28	17.99 ± 1.2	18.74 ± 1.21	73.72 ± 0.71
Burnt
Range	25.08–83.67	0.28–2.26	3.01–8.91	3.05–9.01	66.73–86.59
AVG	75.77 ± 14.29	0.93 ± 0.43	5.78 ± 1.45	5.86 ± 1.45	80.81 ± 4.14

3.2. Total mercury (Hg)

Mercury content in the analysed samples showed an average of 17.1±19.46 ng g^− 1 (n = 21; range from 1.25–63.29 ng g^− 1; unburnt: 10.20 ± 4.22 ng g^− 1, burnt: 19.25 ± 21.18 ng g^− 1). The distribution of mercury concentration per species is shown in Fig. 6. No statistically significant differences were found in mercury concentration among samples by bone type, species, distribution in the cairn (i.e., pyre layers), or burning degree. In all categories of each animal species, a few samples show low mercury concentrations, but the samples with higher concentrations show a decrease in values as m-Stiner grade (i.e., temperature) increases (Fig. 6).

3.3. FTIR-ATR spectra and indices

The spectra (Fig. 7) of the 107 samples show high absorbance in the region 1121 − 900 cm^− 1, which corresponds to ν₁, ν₃ phosphate (PO₄³⁻ asymmetric stretching). Moderate-to-low absorbance is observed at ~ 2010 cm^− 1 (cyanamide), 1680 − 1590 cm^− 1 (Amide I), 1490 − 1360 cm^− 1 (carbonate, CO₃²⁻ asymmetric stretching and CH₂ wagging + bending), ~ 630 cm^− 1 (OH (hydroxyapatite) and 611 − 525 cm^− 1 (ν₄PO₄³⁻ doublet). Low absorbance is shown in ~ 3570 cm^− 1 (OH hydroxyapatite), 1590 − 1500 cm^− 1 (amide II), ~ 872 cm^− 1 (carbonate, (CO₃²⁻) symmetric bending), ~ 698 cm^− 1 (NCN deformation of cyanamide; Snoeck et al. 2014a, Marques et al. 2021). The absorbance bands assignation is shown in Table 3).

Table 3
Selected bands of the spectra and its molecular assignation
Band (cm^− 1)	Assignation	References
~ 3570	νOH⁻	Thompson et al. 2013
3500 − 3030	NH stretching (Amides A-B) – OH⁻	Figueiredo et al. 2012
2982 − 2900	Aliphatics (CH₂ stretching)	Figueiredo et al. 2012
~ 2010	C ≡ N stretching (cyanamide)	Marques et al. 2021; Gonçalves et al. 2023
1680 − 1590	C = O stretching (Amide I)	Figueiredo et al. 2012
1590 − 1500	CN stretching and NH bending (Amide II)	Figueiredo et al. 2012
1490 − 1360	ν₃ CO₃²⁻	Figueiredo et al. 2012
1121 − 970	ν₃PO₄³⁻	Figueiredo et al. 2012; Thompson et al. 2013
~ 961	ν₁PO₄³⁻	Figueiredo et al. 2012; Thompson et al. 2013
~ 872	ν₂ CO₃²⁻	Thompson et al. 2013
~ 698	N-C ≡ N bending (cyanamide)	Snoeck et al. 2014a; Marques et al. 2021
~ 630	OH⁻ libration	Thompson et al. 2013;
~ 598	ν₄PO₄³⁻	Figueiredo et al. 2012; Thompson et al. 2013
~ 560	ν₄PO₄³⁻	Figueiredo et al. 2012; Thompson et al. 2013

Following m-Stiner classification, some differences in the spectra were observed among samples. A description of these spectra as well as the plots of the spectra are available in supplementary text 1, and supplementary Fig. 2, respectively.

Correlation coefficients were calculated between the absorbance intensity of the selected bands and all CIELab parameters, but only considering the burnt group (supplementary Table 3) since it accounts for the vast majority of the 21 samples selected for analytical colour – when projected in the CIELab colour components’ space they group in two different populations (Fig. 5). Correlation coefficients are shown in supplementary Table 3. Lightness (L*) is positively correlated to carbonate absorbances (1495, 1454, 1507, 870 cm^− 1) and negatively correlated to amide I (~ 1620 cm^− 1), cyanamide (2028, 2001, 693 cm^− 1), and phosphate (1075, 598, 556 cm^− 1) absorbances, although correlations are moderate to weak. a* is positively correlated to amide (1690 − 1600, 1570 − 1490 cm^− 1), ν₃CO₃²⁻ (1460 − 1310 cm^− 1) phosphate (~ 1145 cm^− 1) and 616 cm^− 1 absorbances and negatively correlated to cyanamide (2020, 698 cm^− 1), and phosphate (1085, 961, 602 cm^− 1) absorbances; while b* is positively correlated to carbonate absorbances (1575 − 1310, 872 cm^− 1), ν₃ phosphates (1180 − 1100, ~ 1074 cm^− 1), mineral inclusions (770 − 710 cm^− 1), 615 and 570 cm^− 1 (OH_lib valley and ν₄ phosphate valley, respectively) and negatively correlated to OH⁻ (3560, ~ 641 cm^− 1), Amide A (3248 cm^− 1), cyanamide (~ 2028, 2001, 698 cm^− 1), and phosphate (961, 571, 557 cm^− 1) absorbances. C* presents positive correlations with carbonates and amide II (1575 − 1310, 872 cm^− 1), phosphate (1180 − 1100, ~ 1074 cm^− 1), mineral inclusions (770 − 710 cm^− 1), 615 and 570 cm^− 1 (OH(Hap) valley and ν₄ phosphate valley, respectively) and negative correlations with OH(Hap) (~ 3560, ~ 641 cm^− 1), Amide A (~ 3248 cm^− 1), cyanamide (2028, 2001, 698 cm^− 1), and phosphate (961, 571, 557 cm^− 1) absorbances; and h° presents positive correlations with OH (Hap) (~ 3560, ~ 650 cm^− 1), amide A ((~ 3248 cm^− 1), amides I and II (~ 1652, 1560 cm^− 1), carbonates (1490 − 1390, 872 cm^− 1), and phosphate (1110 − 1030, ~ 961, 602, 598, 571 cm^− 1) absorbances.

3.3.3. Mercury (Hg) and FTIR

For the burnt samples (m-Stiner grade > 0) Hg is positively correlated with bands at 3578, 1561, 1525, 1246, 1145, 1028, 720, 621, 565 cm^− 1 and negatively correlated with bands at 2010, 1417, 1104, 1066, 974, 961, 871, 700, 646, 609, 591, 552 cm^− 1. (See Table 4).

Table 4
Spearman correlation coefficient (r_s) for Hg concentration (with the Napierian logarithm correction) for the main spectroscopic IR bands
WN (cm^− 1)	553	566	592	622	646	700	717	873	962	1029	1103	1205	1285	1562	2011
r_s	-0.46	0.62	-0.46	0.16	-0.24	-0.18	0.17	-0.31	-0.33	0.48	-0.36	0.51	0.53	0.47	-0.18

3.4. MIR-indices

Regarding the MIR-indices, the average values and the ranges are displayed in Table 5. Two patterns are observed for most of the indices: i) there is a trend towards increasing or decreasing with m-Stiner grade that shows a distinctive change between grades 3 and 4. Examples of this trend are shown in Fig. 8 and all indices can be found in Supplementary Fig. 3. The CO/CO₃, PHT and FWHM indices are an exception to these trends. As for OH/P, this index is of no application for unburnt samples, since the OH⁻ _lib band does not appear at temperatures below 700°C (Gonçalves et al. 2018). However, from grade 4 onwards, a slight increasing trend of the index is shown. Also, CO/CO₃ experiences a sharp decrease from grade 0 to grade 2 and keeps decreasing after this. FWHM does not seem to follow any trend regarding temperature.

Table 5
Maximum, minimum, average and standard deviation for the MIR-indices.
	IRSF	BPI	C/C	OH/P^a	C/P	API	CN/P	CO/CO₃	Am/P	PHT	_TPV	FWHM
MAX	7.44	0.38	1.94	0.56	0.24	0.29	0.13	1.31	0.24	1.85	1.73	128.92
MIN	2.83	0.02	0.86	0.23	0.01	0.01	0.00	0.08	0.00	0.98	1.02	43.22
AVG	4.88	0.11	1.28	0.40	0.06	0.05	0.02	0.40	0.03	1.35	1.27	77.96
STD	1.00	0.08	0.22	0.06	0.05	0.07	0.02	0.29	0.05	0.23	0.16	18.48

^aTemperature-related indices would be OH/P, CN/P, PHT, CO/CO3 as well as API (because the amides II also present absorbance at the band at ~ 1540 cm^− 1, so in samples with collagen content, this index would be reflecting amides + carbonate type A/phosphate).

3.5. Principal component analysis (PCA, PCA_N, PCA_C) and Discriminant analysis (DA)

The first PCA applied to all the samples extracted five principal components that accounted for 92.3% of the total variance. PC1 (32.0% of the variance) shows high positive loadings for carbonates and amides and high-to-moderate negative loadings for phosphate, hydroxyl groups and cyanamide. PC2 (18.0% of the variance) shows moderate positive loadings for ν₃ phosphate and moderate negative loadings for 1086, 961 and 872 cm^− 1. PC3 (16.4% of the variance) shows moderate positive loadings for 629 and 1086 cm^− 1 and moderate negative loadings for cyanamide and 961 cm^− 1. PC4 (14.5% of the variance) shows moderate positive loadings for 598, 961, and 1455 cm^− 1 and moderate negative loading for 1086 cm^− 1. PC5 (11.4% of the total variance) shows moderate negative loading for 1455 cm^− 1. Loadings for each PC are shown in Supplementary Table 4. Scores for the PCs are projected in Supplementary Fig. 4. The PC1/PC2 projection clearly shows that samples distribute in two different populations (Fig. 9): one with PC1 positive scores and another with slightly positive or negative PC1 scores. The first group shows low to very low variation in PC2 scores, while the second group shows both positive and negative PC2 scores. The first group comprises samples classified as m-Stiner grades 0, 2, 3 and some of grade 4, and the other group encompasses samples classified as m-Stiner grades 5, 6A, 6B and some of the 4, which is consistent with the distinctive change shown by the MIR indices between grades 3 and 4.

To check for the consistency of these two groups, we performed a discriminant analysis, in stepwise mode, assigning samples to each group based on the PC1 scores’ values (supplementary Fig. 5). One discriminant function, CF1, enabled a perfect (100% correct classification) discrimination of the two groups of samples. Seven absorbances were selected for the discrimination, four with positive standardized coefficients (1560, 1455, 962, and 599 cm^− 1) and three with negative values (3344, 1652, and 1417 cm^− 1). Samples with positive PC1 values showed negative CF1 values and samples with negative PC1 scores showed positive CF1 score values. Thus, the DA results support the existence of two clearly different groups: one composed by samples with m-Stiner grade 0–3(4), that correspond to non-calcined samples (T < 700°C), and another with grades (4)5–6(A and B), that corresponds to calcined (T > 700°C) samples. Because of the consistency of these results, we performed a new PCA for each group separately, and the correlation between MIR-indices and PC’s were perform as separate groups as well.

PCA of non-calcined group (PCA_N) resulted in five PCs that accounted for 96.4% of the total variance. PC1_N (32.1% of the total variance) shows high-to-moderate positive loadings for amide I, A and II, ν₂CO₃²⁻ and ν₁PO₄³⁻ and high negative loading for ν₃PO₄³⁻. PC2_N (28.1% of the variance) shows high-to-moderate positive loadings for ν₂ and ν₃ CO₃²⁻ and moderate negative loading for ν₃PO₄³⁻. PC3_N (15.2% of the total variance) shows high-to-moderate positive loadings for 961 and 872 cm^− 1 bands and moderate negative loadings for ν₁ and ν₄ PO₄³⁻ (1015 and 560 cm^− 1). PC4_N (11.2% of the total variance) shows high positive loading for 1088 cm^− 1 and moderate negative loading for 560 cm^− 1 (ν₄PO₄³⁻). PC5_N (9.7% of the total variance) shows high positive loading for 598 cm^− 1 ((ν₄PO₄³⁻). Loadings for each PC_N are shown in Supplementary Table 5.

PCA of calcined group (PCA_C) extracted six PCs that accounted for 91.4% of the total variance. PC1_C (22.4% of the variance) shows high-to-moderate positive loadings for ν₄, ν₃PO₄³⁻ 3571 cm^− 1 (OH⁻ stretching) and amide II and high negative loadings for ν₂CO₃²⁻ and ν₁PO₄³⁻. PC2_C (20.0% of the variance) shows high positive loadings for ν₃CO₃²⁻, moderate loading for ν₂CO₃²⁻ and a moderate negative loading for 698 cm^− 1 and OH⁻. PC3_C (18.1% of the variance) shows high-to-moderate positive loadings for ν₄PO₄³⁻ and OH⁻ and high negative loading for 698 cm^− 1. PC4_C (13.4% of the variance) shows high positive loading for 629 cm^− 1 (OH⁻) and high-to-moderate negative loading for cyanamide (2012, 698 cm^− 1). PC5_C (9.1% of the variance) shows high positive loading for amide II (1560 cm^− 1) and moderate negative loading for 3571 cm^− 1 (OH⁻). PC6_C (8.4% of the variance) shows high positive loading for 1086 cm^− 1 and moderate negative loading for 961 cm^− 1 (ν₁PO₄³⁻). Loadings for each PC_C are shown in Supplementary Table 6.

3.5. Spectroscopic differences between groups

Non-calcined group. The Kruskal-Wallis test showed that IRSF is significantly higher (p = 0.043) for horse bones compared with dog. Also, PC3_N is significantly higher (0.048) in bones with m-Stiner grade 3 than in bones with m-Stiner grade 0, and CO/CO₃ is significantly higher (p = 0.038) in samples classified as grade 4 than in samples classified as grade 0 (see Fig. 10).

Regarding the indices, IRSF is negatively correlated to BPI, C/P, Am/P and FWHM (ρ = -0.783; p < 0.001, ρ = -0.805; p < 0.001, ρ = -0.732; p < 0.001, ρ = -0.761; p < 0.001, respectively). BPI is positively correlated to C/P (ρ = 0.974; p < 0.001). Also, FWHM is positively correlated to BPI, C/P and N/P (ρ = 0.761; p < 0.001, ρ = 0.829; p < 0.001, ρ = 0.914; p < 0.001) and. The MIR indices also showed significant correlations to the extracted principal components: PC1_N is negatively correlated with IRSF (ρ = -0.554; p = 0.006) and positively correlated with Am/P (ρ = 0.880; p < 0.001) and FWHM (ρ = 0.745; p < 0.001). PC2_N is positively correlated with BPI ((ρ = 0.849; p < 0.001) and C/P ((ρ = 0.779; p < 0.001). See supplementary Fig. 7.

Calcined group. The Kruskal-Wallis test showed that, for the pyre layers OH/P and CO/CO₃ are significantly higher in samples from the Badger layer than those of layers S3 and S2 (p < 0.001) and Am/P is significantly higher in the Badger layer than in S3 samples (p < 0.001). Also, some differences were found among bone types: OH/P is significantly higher in thoracic bones compared with hand/foot bones (p = 0.002) and cranial bones (p = 0.002). CN/P is also higher in hand/foot than in crania (p = 0.002), and PHT is higher in thoracic bones compared with crania samples (p = 0.008). Regarding m-Stiner classification statistically significant differences were found between grades 6A and 6B: samples classified as 6B showed higher C/P values than samples from the 6A group (Fig. 11).

Regarding correlation to the MIR indices, IRSF is negatively correlated to BPI (ρ = -0.606; p < 0.001), and C/P (ρ = -0.677; p < 0.001), and positively correlated to C/C (ρ = 0.479; p < 0.001), PHT (ρ = 0.574; p < 0.001), _TPV (ρ = 0.751; p < 0.001). BPI is positively correlated to C/P (ρ = 0.970; p < 0.001) and negatively correlated to PHT (ρ = -0.546; p < 0.001), _TPV (ρ = -0.635; p < 0.001). PHT is positively correlated with _TPV (ρ = 0.875; p < 0.001) and negatively correlated with C/P (ρ = -0.530; p < 0.001). C/P is negatively correlated with _TPV (ρ = -0.640; p < 0.001). As for the correlation to the principal components, PC1_C is positively correlated to IRSF (ρ = 0.779; p < 0.001), PHT (ρ = 0.317; p = 0.003), _TPV (ρ = 0.503; p < 0.001) and negatively correlated to BPI (ρ = -0.320; p = 0.003), C/P (ρ = -0.461; p < 0.001) and FWHM (ρ = 0.907; p < 0.001). PC2_C is positively correlated to BPI (ρ = 0.883; p < 0.001) and C/P (ρ = 0.814; p < 0.001) and negatively correlated to CO/CO₃ (ρ = -0.792; p < 0.001) and PHT (ρ = -0.497; p < 0.001). PC3_C is positively correlated with PHT (ρ = 0.469; p < 0.001). PC4_C is negatively correlated to CN/P (ρ = -0.626; p < 0.001) and positively correlated to OH/P (ρ = 0.821; p < 0.001). See supplementary Fig. 8.

4.1. Mercury content in bones

The detection of measurable mercury concentrations in cremated bone is somewhat surprising, even in those samples that were exposed at high temperatures, i.e., calcined. Elemental mercury volatilizes at relatively low temperatures (Hg° is released at 150°), while other mercury species are retained until higher temperatures (HgS gets mainly released at 400°) (Carvalhinho Windmöller et al. 2017). In our samples, we have observed a general trend towards a decrease of mercury concentration and variability with increasing temperature. Therefore, there is a temperature-effect in mercury concentration. This agrees with the observed Hg association to the presence of OH⁻ (3578, 621 cm^− 1), collagen bands such as amide II (1561, 1525 cm^− 1) and III (1246 cm^− 1), ν₁PO₄³⁻ (1028 cm^− 1) and with its decrease in more crystalline samples (high 961, 591, 552 cm^− 1.absorbances) and samples with cyanamide (2010, 700 cm^− 1), i.e., high temperature. However, despite displaying the human samples the higher burning degree, their mercury concentrations tend to be higher (five samples being over the average) than unburnt dog samples, which display values below average – suggesting a species effect also in mercury concentration. The coexistence of both effects -temperature and animal species- as well as different Hg species, points towards the presence of strong mercury bonds in human remains that allowed for the retention of mercury in bone despite being burnt at low-mid temperatures, and discard any kind of diagenetic incorporation.

Archaeological bone can incorporate mercury during diagenesis when buried in soils rich in mercury (such as those close to mercury mines or formed from parent materials rich in this heavy metal), but this is not the case of Swedish soils. If mercury came from a modern contamination of topsoils (Panagos et al. 2021), no differences between species would exist. Thus, diagenetic incorporation is not likely in this case. In a previous investigation from our team (Álvarez-Fernández et al. 2022), we observed an inverse relationship between mercury and bioapatite crystallinity in inhumated skeletons. However, in the present study, we did not observe any correlation between IRSF and mercury concentration although we did observe a temperature effect - the higher the temperature, the lower mercury concentration in bone is. These two observations led us to discard a post-mortem diagenetic incorporation of mercury in our samples. Also, in the aforementioned investigation (Álvarez-Fernández et al. 2022) our team found that mercury was sourced from two different processes: i) ante-mortem incorporation during life, and ii) post-mortem bonding during soft tissues decay, especially from kidneys and liver (target organs for mercury intoxication) (García et al. 2001) leading to an intra-skeletal effect. Usually, in cremation funerary practices, the deceased individuals are burnt before the decomposition of soft tissues takes place. However, the incomplete combustion of the organic fraction under low burning temperatures observed in some of the samples (French et al. 2022), could have led to the retention of mercury in the carbon deposited during the cremation.

Carbon atoms in cremated bones can also come from exogenous sources such as the wood from the pyre (Zazzo et al. 2012). However, the observed differences in mercury concentration between species cannot be explained by mercury from the deposited carbon from the wood of the pyre. The presence of mercury could also be related to the degree of oxygen availability during burning, which would explain its positive correlation with OH⁻ and negative correlation with cyanamide. Mercury concentration could be higher in samples burnt under oxidizing conditions and lower in samples burnt under reducing conditions, a fact that can be linked to the highest concentrations in crania, fibula and humerus, areas burnt more quickly than thoracic bones. More systematic studies sampling various body locations with the same degree of burning would be necessary to test this hypothesis.

Based on the results presented here and comparison with previous studies, we interpret that the mercury incorporation happened during life instead of coming from post-mortem alteration. However, when compared to other studied sites referred in the literature (see table in Álvarez-Fernández et al. 2020), the observed Hg concentrations are lower (17.1 ± 19.46 ng.g^− 1) than the mercury levels documented for inhumated human remains from rural Medieval Denmark (Rasmussen et al. 2008; Rasmussen et al. 2017; Rasmussen et al. 2015) or the post-Roman A Lanzada in Spain (Álvarez-Fernández et al. 2020; López-Costas et al. 2020). The lack of coetaneous studies on inhumated remains in Sweden makes it impossible to determine whether the low levels are due to the effect of fire or to a low exposure to mercury pollution. In all cases, the low detection limit of the triple-cell DMA equipment used here opens the door to exploring ante-mortem mercury pollution through the study of cremated bones and its potential to be used as a proxy of the temperature achieved during burning.

4.2. Colour as temperature indicator

Most bones in this study fall into m-Stiner grade 6 (A and B), which could indicate either a good burning of the samples or it may respond to the fact that calcined bone is less susceptible to diagenesis, since the HAp crystals are bigger and more stable (Gallo et al. 2021). It is important to bear in mind that changes between m-Stiner grades of burning for this collection may be reflecting not only structural differences originated by the burning process at different temperatures and time of exposure, but also by differential diagenetic alteration since samples from grade 3 were more prone to diagenesis than samples from grade 6. Walker et al. (2008) pointed out that factors such as presence of organic compounds and oxygen availability have a greater influence in colour changes than time of exposure to fire.

We detected that macroscopic observation, in spite of being informative, non-destructive and affordable (Evans et al. 2022), does not completely match the information obtained by IR spectroscopy. Part of this disagreement may be explained by the inter-observer effect, and subjectivity of the macroscopic analysis (Krap et al. 2017). The differences we observed in the macroscopic m-Stiner assignation, and the spectroscopic data may be related to the fact that changes in colour are not only determined by the heating temperature but also by the time bone is exposed to heat (Greiner et al. 2019). Complete loss of organics can occur at ~ 300°C and calcination can start at ~ 550°C with long-exposure heating times (Gallo et al. 2023). However, this does not seem to be the case for our samples, since the relationship of colour and IR parameters, when looking at m-Stiner classification, and the MIR-indices and PCs although not being perfect it is relatively strong, and it is even improved when comparing IR values with analytical colour values.

Analytical colour analyses in the CIELab space have proven to be reliable temperature indicators before (Krap et al. 2019; Rubio et al. 2020). In our study, we analysed bones from different species and probably subjected to higher temperatures (m-Stiner grade 6B; >900ºC). Like Rubio et al. (2020), we observed that L* decreases with temperature until 400°C and then it increases again. We observed a new slight decrease at m-Stiner grade 6B that Rubio et al. (2020) did not observe, but these authors only studied samples heated at a maximum temperature of 800°C. Also, those authors observed that a* and b*were higher for samples heated at temperatures lower than 400°C and then they decreased drastically, which agrees with our results. Paba et al. (2021) also observed a slight decrease of L* from 900ºC to 1000ºC. Krap et al. (2019; 2024) proved that L* and b* values have high potential for estimating burning temperature even in faunal remains and embalmed human remains.

Walker et al. (2008) found that, to some extent, analytical colour measured in the RGB system was a good predictor of collagen content. This agrees with our results since we have found that collagen bands correlate with colour parameters: for samples with observable 1600, 1560, 1524 cm^− 1 intensities the peak absorbance correlates negatively with L*, and positively with a* and b*, indicating that samples with collagen would be redder, yellower, and darker. This opens the door for using analytical colour as a predictor for collagen presence in bones. However, Gallo et al. (2023) observed that samples heated at 300°C during prolonged time exposure (48 hours), did not achieve calcination even though the samples presented pale colour, thereby, caution must be taken when inferring temperature by bone coloration. Other factors, such as the presence of iron and manganese oxides, may also produce dark colours even at high-temperature exposure (Michel et al. 2006). However, it does not seem the case for our samples, since the spectroscopic signal and the analytical colour values display good correlation.

4.3. Interpreting the PCA results

For the studied samples, PC1_N represents collagen vs. phosphate content; PC2_N represents carbonate abundance relative to phosphate; PC3_N represents the relative intensity of 961 and 872 cm^− 1 (temperature-related bands) versus 560 cm^− 1; PC4_N represents the relative intensity of 1088 versus 560 cm^− 1 and PC5_N represents the relative intensity of the 598 cm^− 1 band. The scores are projected in Supplementary Fig. 5. Even though we have classified these samples as a group, scores for PC1_N show that bone samples with m-Stiner grade 0 have higher values compared to those classified as m-Stiner grades 2, 3, 4 and 5, which indicates that burning affects collagen content even at low temperatures. Also, human samples show relatively higher positive scores for PC2_N, indicating that this group may present higher carbonate-to-phosphate content than horse and dog samples. Most of the samples have negative scores for PC3_N except for Tyr094, Tyr091, Tyr101, Tyr105, Tyr013, Tyr082, Tyr092, Tyr096, Tyr050 and Tyr107. PC4_N shows large positive scores for Tyr101 and Tyr102 and large negative scores for Tyr091 and Tyr013. PC5_N shows especially high positive scores for Tyr085 and Tyr050 and high negative scores for Tyr101 and Tyr104.

Regarding PCA_C, the results indicate that, PC1_C represents ν₄, ν₃PO₄³⁻ and amide versus ν₂CO₃²⁻ and ν₁PO₄³⁻. PC2_C represents carbonate content versus 698 cm^− 1 and OH⁻. PC3_C shows ν₄PO₄³⁻ and OH⁻ to 698 cm^− 1. PC4_C shows the ratio of OH⁻ versus cyanamide and 698 cm^− 1. PC5_C represents A-type carbonate-to-OH⁻, and PC6_C shows 1086 cm^− 1 versus 961 cm^− 1. The scores for PCA_C are projected in supplementary Fig. 6. For PC1_C most samples classified as 6B show negative scores while samples classified as 6A and 5 show mainly positive scores for this component. The bands involved in this component are temperature-related, which indicates that samples classified as m-Stiner grade 6B experienced a reverse chemical transformation of the phosphates and carbonate components of the hydroxyapatite compared with those classified as 6A and 5, that presumingly achieved lower temperatures than the former. For PC2_C 6A samples present mainly negative scores and 6B samples positives scores, indicating that the chemical changes diagnostic of burning (hydroxyl substitutions and cyanamide formation) are more intense in samples classified as m-Stiner grade 6A than in samples classified as 6B, pointing again towards the inversion of the structural changes in bone with heating. For PC3_C most samples present positive loadings, but samples Tyr024, Tyr025, Tyr026, Tyr029 Tyr045, Tyr051, Tyr073 and Tyr099 present high negative loadings, indicating that those samples are less crystalline than the rest. For PC4_C, most samples present positive scores except for Tyr011, Tyr014, Tyr040, Tyr055, Tyr068, Tyr071, Tyr073, and Tyr084, all of them are samples from hand/foot, except for a patella and a long bone fragment (see Fig. 12). This component indicates cyanamide formation, which points towards a relation between bone type and the presence of this compound. For PC5_C samples Tyr018, Tyr026, Tyr036 present considerably higher positive loadings than the rest, and Tyr020, Tyr024 and Tyr029 present higher negative loadings than the other samples, indicating that the OH⁻ substitutions of A-type carbonate were not as extensive in the former compared with the latter. For PC6_C half of the samples present positive loadings, and this seems to respond to a distribution pattern: samples from layers S3 and ÖB show mainly negative loadings and samples from Badger layer seem to have mainly positive loadings for this component. The band at ca. 1086 cm^− 1 involved in this component is characteristic of calcined samples, but it has not been thoroughly studied yet, and we discuss it below.

4.4. Transformation of bone with temperature: shift between m-Stiner grades 3 and 4.

In this study, we observed that the differences in the spectroscopical signal between calcined samples and non-calcined samples are larger than the ones among samples within those groups, even between different m-Stiner burning grades. This differentiation was already somewhat observed by Gallo et al. (2021) who found that IRSF allowed to distinguish between calcined and not calcined samples. This shift in bone structure between 400ºC and 700ºC was also observed in previous studies (Etok et al. 2007; Galeano and García-Lorenzo 2014) in which cremated bone samples were analysed by XRD, showing an increase in crystals size. Gonçalvez et al. (2018) also found substantial changes in IRSF, BPI and OH/P between 700–800°C. These authors found that IRSF keeps slightly increasing after 800ºC. However, other authors (Thompson et al. 2013; Snoeck et al. 2014a; Ellingham et al. 2016), have described a slight decrease of IRSF values (and thereby crystallinity) with temperature after reaching 700–800°C. This is also shown in our case, since IRSF values (as well as the general slight change of trend observed in other MIR-indices) are lower for m-Stiner grade 6B than grade 6A samples. Ellingham et al. (2016) attributed the decrease in crystallinity (and decrease in C/C, PHT, increase in CO₃/P, CO/CO₃) to the sintering process, and formation of β-TCP.

Those abrupt changes in crystallinity were also observed by Greiner et al. (2019), who attributed them to the loss of organics that no longer coat the apatite crystals, allowing their growth. The strong shift in the indices’ values - and the separation in two groups of samples based on their IR-spectroscopic signals - seems to respond to the complete elimination of organic contents which leaves the inorganic part exposed and, therefore, changes in its structure occur (Person et al. 1996; Thompson et al. 2013). These structural changes give the calcined bone completely different properties than bones that underwent diagenesis or low temperature exposure (Lebon et al. 2010). At which temperature this change occurs is not clear yet. According to Etok et al. (2007), proteins are removed at temperatures of 400°C while other organic components (such as aromatic compounds) are still present up to 600°C; although, Gallo et al. (2023) observed that this can occur under lower temperatures at long-exposure times. However, in our case, the aromatic signal was only present in Tyr076 as a subtle shoulder. This sample is a dog carpus classified as m-Stiner grade 6A. Also, some of the organics’ absorbances are shown in most of the samples, even in some of the samples classified as 6A and 6B, in which a slight absorbance at the amides I and II regions is observed. However, some samples lack organics completely, especially those from the 6A grade. This agrees with the findings of Mamede et al. (2018b), as they did not observe a complete loss of organics until temperatures higher than 700°C, but they attributed this phenomenon not only to the temperatures reached but also to the duration of burning.

In addition to the removal of organic compounds, carbonate ions substitutions are also involved in the increase of apatite crystallinity. This is shown in our samples by the negative correlation between IRSF and C/P, and BPI, but positive correlation with C/C. This was documented in previous studies (Snoeck et al. 2014b; Végh et al. 2022) and indicates that crystallinity increases as carbonate content relative to phosphate decreases (loss of carbonate ions). However, in our study, the C/C, i.e., A-type carbonate substitutions relative to B-type carbonate substitution, tends to increase with temperature once calcination is reached, even though the overall amount of carbonate decreases, in agreement with previous studies (Snoeck et al. 2014a; Gonçalves et al. 2018; Brandao et al. 2023). Marques et al. (2018) stated that B-type carbonate is more sensitive to heat than A-type carbonate and starts to decrease earlier, which is coherent with our results, since C/C increases from m-Stiner grade 4 onwards. Another factor to consider is the substitution of some hydroxyl groups by carbonate A-type at high burning temperatures (Zazzo et al. 2012; Thompson et al. 2013). However, care must be taken when interpreting C/C changes, since this index also reflects the presence of lipidic signal at 1450 cm^− 1 in bones burnt under temperatures of 800°C (Gonçalves et al. 2018). These organic compounds are the first in disappearing with temperature (Marques et al. 2018), thereby, it is possible that the initial decrease of this index from m-Stiner grades 0 to 3 may also reflect the loss of lipidic content. We have also observed that some of the samples of m-Stiner grade 4 onwards, present a double peak at ν₂CO₃⁻² region (872 cm^− 1 and 879 cm^− 1), and part of the highly calcined samples had even completely lost the band at 872 cm^− 1. This phenomenon has been attributed to A-type substitution of the OH⁻ group and has been described to occur at temperatures higher than 700°C (Fleet 2009; Thompson et al. 2013; Thompson et al. 2013; Greiner et al. 2019) (see. Figure 13). Thus, we think that during calcination, global carbonate content decreases, but some A-type carbonate substitutes in OH⁻ sites of the hydroxyapatite, resulting in an increase of the C/C index (more A-type carbonate relative to B-type carbonate) and the appearance of the ~ 879 cm^− 1 band. This may be due to a secondary carbonate formation from the CO₂ released during cremation (Hüls et al. 2010; Gonçalves et al. 2018; Brandao et al. 2023).

The bands at 3570 and 630 cm^− 1, which have been suggested as indicators of calcination (e.g., Gallo et al. 2023), appear because of the loss of A-type carbonate ions and its substitution by hydroxyl ions (OH⁻) (Etok et al. 2007; Snoeck et al. 2014a; Mamede et al. 2018a; Shaw 2022). This is coherent with the negative correlation between PHT versus BPI and C/P observed in our samples. Shaw (2022) found that, at least for faunal remains, the 630 cm^− 1 band never appears at temperatures under 525–537°C even during long-exposure times. In our samples, PHT is positively correlated to IRSF, which indicates an increase of the OH⁻ libration band with crystallinity, which according to Snoeck et al. (2014b) is due to the fact that hydroxyl ions are smaller than carbonate ions, allowing for the apatite structure to get more crystallized. PHT is also positively correlated to _TPV, which we also attributed to an increase in crystallinity, and we discussed it below.

Among the calculated MIR-indices, OH/P, CO/CO₃ and FWHM do not respond to the abrupt increase/decrease pattern from m-Stiner grade 3 to grade 4. FWHM, in particular, does not seem to respond to temperature changes in our samples, disagreeing with previous studies. For example, Thompson et al. (2013) observed that FWHM decreased with temperature until the 800ºC, but when reached, it grew again, although not as much as in unburned bone.

4.5. Transformation with temperature: formation of new bands: 2012, 1088, 961, 699 cm^− 1

Some new absorbance bands are shown in the spectrum of bones at ~ 2012 and 699 cm^− 1 (cyanamide) and at 1088 and 961 cm^− 1 (ν_1,3 PO₃ ²⁻) over temperature.

4.5.1. ν_1,3 PO₃ ²⁻ bands (1088, 961 cm^− 1)

Both bands have been shown to appear and sharpen with increasing m-Stiner grades. In particular, the band at 961 cm^− 1 has been found to also appear and sharpen in inhumated bone subjected to severe diagenesis (see for example spectra in Lebon et al. (2016) and Castorina et al. (2023)). This goes in accordance with Querido et al. (2018), who observed that absorbance at 960 cm^− 1 is correlated to the crystallinity of biological and synthetic apatites. The index proposed here, _TPV, is positively correlated with IRSF. This may indicate a relation between the sharpening of 1088 cm^− 1 with the increase in crystallinity of the bioapatite. Also, _TPV is positively correlated with PHT and negatively correlated with BPI. In addition, the PCA_C, suggest that this band covariates positively with OH⁻ (630 and 3572 cm^− 1) and negatively with amide II (see Supplementary Table 6). Thus, it is possible that the loss of carbonate ions and their substitution by hydroxyl anions plays a role in the sharpening of 1088 cm^− 1 peak. Despite this, Reidsma et al. (2016) found this band as a subtle shoulder in samples heated at 900°C under reducing conditions, thereby, the incorporation of hydroxyl ions may not be the cause of its appearance, or perhaps other factors are involved in its formation. The _TPV index starts increasing at samples classified as m-Stiner grade 4 (samples start to get calcined), and keeps increasing at grade 5, grade 6A and then decreases slightly at grade 6B (see supplementary Fig. 3). This indicates that the _TPV is a promising index for inferring bone temperature, and it would be interesting to explore it in experimental studies. This agrees with Gallo et al. (2023), who used the presence/absence of the 1088 cm^− 1 band as a proxy of temperature transformations, in particular calcination. Some authors (Marques et al. 2018; Festa et al. 2019) consider 1088 cm^− 1 band as indicative of the transformation of the mineral lattice into fluorapatite (francolite; Ca₅(PO4)₃F). In agreement with this hypothesis, previous studies (Geiger and Weiner 1993; Trueman et al. 2008) have described higher 603 cm^− 1 intensities when F⁻ substitutes in the OH⁻ sites of the bioapatite. This phenomenon would increase the IRSF values and therefore would explain the correlation between this index and the 1088 cm^− 1 peak. Also, in our samples, a higher absorbance of 599 cm^− 1 relative to 560 cm^− 1 is observed and is consistent with the PCA_C results in which 560 cm^− 1 band is grouped in PC1_C with 1023 cm^− 1 band, but the band at 599 cm^− 1 is grouped in PC3_C with 3572 cm^− 1 band (hydroxyls, thus it is temperature-related). Marques et al. (2018) explain the absence of the 1088 cm^− 1 band at low temperatures as being obscured by the main ν₃PO₄³⁻ band, which would become narrower at higher temperatures. However, we have not observed any tendency for the broadening of this band with temperature and the _TPV index does not correlate with FWHM (Fig. 7). Festa et al. (2019) also interpreted the band at ~ 961 cm^− 1, together with ~ 1086 cm^− 1, to be indicative of fluorapatite. However, in our samples, these two bands are not grouped in the same PC in the PCA_C (i.e., they do not covary). Lebon et al. (2010) observed that archaeological samples present Sr^+ 2 and F⁻ intakes. However, they noticed that calcined samples did not have a highly altered elemental composition, and we observe in our samples that the band at ~ 1088 cm^− 1 sharpens with increasing temperature (as observed by Berna 2010), which means that F⁻ incorporation is probably not the cause (or at least not the main one) for the presence and sharpening of this band. In addition, Stathopoulou et al. (2008), analysing mammal fossil bones, found that diagenesis can also produce and shape this band, it did not correlate with F⁻ concentration - supporting our hypothesis. The study of the elemental composition would shed some light on the relation between the presence of this band and the formation of fluorapatite.

4.5.2. Cyanamide formation

In some Tyresta samples of the calcined group, bands characteristic of cyanamide (2012, 699 cm^-1) were detected. The cause for the formation of this compound is still under debate. Some authors have argued that it may be related to the presence/absence of oxygen during the burning process (Gonçalves et al. 2023). Reidsma et al. (2016) observed that bone heated at temperatures higher than 700°C, in anaerobic conditions, tends towards the formation of cyanamide, and Gonçalves et al. (2023) noticed that this peak’s intensity increases with temperature. One of Gonçalves et al. (2023) criteria for recognizing bones burnt under anaerobic conditions is dark colour. In our samples, L* is indeed negatively correlated with the cyanamide bands, indicating that low lightness is related to the presence of this compound. Furthermore, a* and b* are also negatively correlated to this band, indicating that a bluer-greener coloration of the bones would be related with anaerobic burning environment. Note also that using naked-eye colour for inferring temperature can lead to biases, since blackish colouring may have its origin in reducing-environment conditions instead of low-temperature heating (Reidsma et al. 2016). The anaerobic conditions hypothesis is consistent with our PCA_C results, in which PC3_C and PC4_C represent OH^- vs. cyanamide substitutions (3572 vs. 699 cm^-1 and 630 vs. 2012 cm^-1, respectively). Being those facts indicative of N-C ≡ N bending substituting at 3570 cm^-1 (OH^- stretching) and C ≡ N stretching substituting for 630 cm^-1 (OH^- libration) and would also explain why Gonçalves et al. (2023) found cyanamide absorbance in samples in which the peak at 3570 cm^-1 band (OH^- stretching) was also observable, while Reidsma et al. (2016) did not. Thus, both PC3_C and PC4_C are likely to be related with oxygen availability in the burning environment: samples with positive values would indicate aerobic environments and samples with negative scores for these PCs would indicate anaerobic burning.

In our samples, even though the PCA_C showed that 2012 cm^− 1 anti-covaries with 630 cm^− 1 and 699 cm^− 1 anti-covaries with 3570 cm^− 1, the four bands are present at the same time in the spectra. This has already been shown in previous studies (Zazzo et al. 2013; Snoeck et al. 2014a), and points to a different explanation other than the reduction/oxidation environment of the combustion process. Leskovar et al. (2025) observed that the time of exposure to heating had an impact on the 2012 cm^− 1 band. These authors found that at temperatures of 900°C the intensity of the 2012 cm^− 1 band decreased with higher exposure times, while the 630 cm^− 1 and the A-type carbonate increased. This could explain the differences of CN/P index in our samples according to bone type, since cyanamide is higher in samples with thin, soft tissue (carpal/tarsal bones, patella) while OH⁻ is higher in thoracic bones, that are surrounded by fat and other soft tissues. We hypothesize that the fat may have played a role as fuel (animal fats were used as fuel for firing lamps since at least Iron Age (Mayyas et al. 2017) and they are still used for biodiesel synthesis (Awad et al. 2014; Kumar et al. 2006)), increasing the burning time at these regions, and therefore leading to the substitutions of carbonate by hydroxyl groups rather than cyanamide. This fuel effect was also demonstrated by the experiment of Dehaan (2012), who burned three human corpses under different conditions and observed that subcutaneous body fat acted as fuel for fire leading to the torso, pelvis and upper legs to burn for longer times than hands, feet, head and lower legs. We assume that, in Tyresta, once started the fire was not manipulated until extinguished, following its natural firing patterns of shorter times of firing in hands and feet and much longer for the torso due to the role of fat as fuel.

Cyanamide has been also associated to the presence/absence of ammonia during the cremation process (Mamede et al. 2018a; Gonçalves et al. 2018). In this case, the absence of surrounding flesh of hand/feet, as well as patella, may have created the proper environment for the formation of this compound (perhaps due to presence/absence of organic matter). Some authors have proposed the combustion fuel (higher N content in flesh than in wood) as the origin of the ammonia that leads to the substitution of hydroxyl and A-type carbonates for C₂NH (Snoeck et al. 2014a). According to Zazzo et al. (2013), the carbon may come from the collagen of the burnt bone or from the wood of the pyre. Another possible explanation is the presence/absence of clothing which can be a source of ammonia (i.e., leather shoes or clothes), and it can as well contribute to creating a more reducing environment, allowing the formation of cyanamide (Salesse et al. 2021). However, this is unlikely to be our case, since our samples belong both to human and animal bones. Also, Salesse et al. (2021) only analysed foot bones, which precludes differences among bone types, meaning that perhaps the presence of cyanamide is explained by time of exposure rather than the presence or absence of shoes.

Another possible explanation is the pugilistic posture - i.e., contraction of the muscles with fire resulting in a characteristic posture with flexed arms, legs and wrists (Symes et al. 2015). - in which both the bones from hand and feet as well as patellae are highly exposed, and they are the first to burn (Symes et al. 2015). However, recent investigations have proposed that the pugilistic posture is not a universal phenomenon when studied in real fire settings instead of crematorium chambers, and sometimes the body reaction to burning is the extension of the limbs away from the torso (Harding et al. 2024). Also, some of the bones that show this band belong to horse and dog, so perhaps even if they got the pugilistic posture, the areas exposed to fire would be different compared to those of humans because of the anatomy of these animals and, again, this does not seem the most likely explanation.

4.6. Bone type and species

In the non-calcined group, no statistically significant differences were found between different bone types. However, structural and chemical differences among bone types have been described before in the literature (Gonçalves et al. 2018; Leskovar et al. 2023; Maurer et al. 2023; Colmenares-Prado et al. 2025). The relatively small number of samples of this group (n = 23) and the fact that there were bones from three different species (human, horse and dog), prevented a proper statistical analysis of this group, and we assume that the differences between species may be larger than differences between skeletal elements.

However, in the calcined group (n = 84), we observed that OH/P and PHT are higher in thoracic bones compared with crania and hands/feet. Those two indices were described to indicate oxygen availability during the burning process (Gonçalves et al. 2023). Thereby, this likely means that thoracic bones were burned under aerobic conditions. On the other hand, CN/P is higher in hands/feet than in crania. PC4_C scores are considerably more negative (i.e., cyanamide signal) for hand/foot bones and patellae and a higher number of samples with positive scores (hydroxyl) in thoracic bones compared with the other bone types. This is consistent with Stamataki et al. (2021), who observed that cyanamide (evaluated by the CN/P index) was more abundant in long bones than in ribs.

Thereby, it seems that the differences in bone structure in our samples are more related with burning patterns than with ante-mortem pre-existent differences. This agrees with Krap et al. (2019) that found that ante-mortem differences in colour (measured in CIELab system) of burnt bone among bone type, age, or sex, disappeared at temperatures above 300°C. Since we found that colour and chemical composition are highly related, it is likely that structure-chemical differences disappear with increasing temperature. Spectroscopic differences among bone types were also not observed in previous investigations on cremated bones (Mamede et al. 2018b), or statistical differences were only found for one index (CN/P, Stamataki et al. 2021).

Regarding species, in the literature there are some examples of chemical differences in spectral properties between human and non-humans in inhumated bones (Wang et al. 2019; Martínez Cortizas and López-Costas 2024). We observed statistically significant differences in IRSF between species in the non-calcined group, but not in the calcined group, which might indicate that the changes of bone (i.e., recrystallization) with temperature are stronger than the ante-mortem structural differences among species. Thereby, despite of being tempting to use spectroscopic data to distinguish animal from human bones in the context of a highly fragmented cremation burial, care must be taken when using MIR-indices to discriminate among species in calcined bones.

4.7. Variations with time of exposure

Hiller et al. (2003) observed that hydroxyapatite crystals’ shape and size changed not only with temperature but also with heating time. Recent experimental studies have also focused on the effect of time of exposure to heat (e.g., Mamede et al. 2018b; Legan et al. 2020; Gallo et al. 2023), and the general findings suggest that some changes in bone structure/composition are related to the time of exposure while others are necessarily triggered by temperature. Snoeck et al. (2014a) found that MIR-indices change with heating time until the bone is fully calcined, but once this happens, significant changes are only observed with temperature. If this was the case for the Tyresta samples, changes in the calcined group would be reflecting time variations in exposure to heating.

Both Gallo et al. (2023) and Leskovar et al. (2025) studies found that at 300°C there is a sudden loss of collagen, that keeps decreasing with time of exposure at this temperature. Carbonate loss also started at this temperature, but it did not increase with time of exposure to this temperature. However, Gallo et al. (2023) observed that IRSF did not change with time at 300°C. This would imply that the loss of organics itself is not enough for the mineral changes to occur with heating. Leskovar et al. (2025) observed that OH⁻ ions could incorporate at the bioapatite lattice even at 300°C with long exposure times, while Gallo et al. (2023) did not observe any OH⁻ libration absorbance at this temperature. In addition, Leskovar et al. (2025) found that the presence of the 879 cm^− 1 A-type carbonate band at these temperatures even at short exposure times, which increased as 872 cm^− 1 B-type carbonate decreased. At 600°C the amide I kept decreasing with exposure until its complete disappearance. Also, at this temperature carbonate decreased and OH⁻ increased with exposure time, while IRSF and the 961 cm^− 1 band continued increasing with a sharp change at 90 and 120 minutes. Gallo et al. (2023) found that at 550°C and 750°C the IRSF increased with time (9 hours). However, after 48 hours at 550°C IRSF, PHT and the sharpening of 1090 cm^− 1 band kept increasing but at 750°C IRSF suffered a decrease with heating time. Leskovar et al. (2025) only observed this decrease at long-time exposure at 900°C and higher temperatures (i.e., 1200°C). Leskovar et al. (2025) noticed that at 900°C the amide I, 1415 cm^− 1, and 872 cm^− 1 bands disappeared almost completely while the 878 cm^− 1 and OH⁻ bands increased at long-time exposure times. At 1200°C all the organics and carbonates (including the 878 cm^− 1 band) disappeared. But they detected that at short heating times at 1200°C cyanamide and the 878 cm^− 1 peak can sometimes be present. According to Leskovar et al. (2025), at 1200°C OH⁻ remained unchanged respect to 900°C. In our case, there is a good correlation between indices at the different m-Stiner grades. Thus, it seems that chemical changes are mainly explained by temperature rather than time of exposure to heating (indicating that probably the time of heating was relatively short). According to Carrol and Smith (2018), prolonged fires need human intervention to avoid extinction, therefore, it is likely that in the Tyresta site, once the fire was started, no effort was made on maintaining it alive, leaving it to follow its own course.

4.8. Archaeological implications in Tyresta site

Temperatures reached. Some of the samples studied here reached the calcination state. Also, no signs of long-term exposure are found, thereby, the calcination probably occurred at ~ 700ºC (Gallo et al. 2021; Gallo et al. 2023). Also, part of our samples reached the china-like appearance characteristic of m-Stiner grade 6B. In addition, the spectroscopic signal indicates that the samples did not reach 900–1000ºC, since carbonate loss was not complete (Mamede et al. 2018a; Gallo et al. 2021). This is also supported by the fact that bands involved in the CO/CO₃ index are present in our samples, in agreement with Thompson et al. (2009) findings.

Funerary practices. Our results point to the fact that once the fire was initiated, no more human intervention happened until it was completely consumed. This is inferred from the presence of cyanamide in hand/feet bone samples, since the natural pattern for these types of bones is to burn more rapidly than torso bones (Dehaan 2012).

Badger layer. We have observed a trend towards higher 1088 cm^− 1 towards 961 cm^− 1 band in badger layer in respect to ÖB and S3. This, combined with the fact that samples from this layer were visually different, suggests that the samples come from another burning event, since the band at 1086 cm^− 1 is temperature-specific. Also, the OH/P index is significantly higher for the badger layer compared with layers S2 and S3, indicating that samples from this layer were subjected to higher temperatures. This is not surprising, since most bone fragments from this layer were classified as m-Stiner grade 6. This layer is named after the first archaeological hypothesis that a badger made a tunnel and removed samples from inside the pyre. However, the differences of the temperature achieved point to a different explanation of the existence of this layer, maybe it was intentional and human-crafted. It would be interesting to radiocarbon-date samples from this layer and from inside the cairn to check whether the ages of the bones of the different layers are consistent or not. Also, the statistically higher values of Am/P and CO/CO₃ indices in bones of the badger layer versus S3 (both indices) and S2 (CO/CO₃), points to a higher protein content. This may point towards a different origin of these bones. However, as indicated before, most pre-burning structural and chemical differences between bones, even at species level, get loss after calcination, so this result must be taken with care.

Mercury. Although not extensively documented, there are some studies describing the use of mercury during the Viking age, like the mercury gilded brooches found at the proto-urban Birka site (Sweden) (Ullén et al. 2021) or the high mercury concentrations in stone artefacts identified as possible touchstones for metal amalgamation in the Viking age burials of Hedeby (Germany) (Ježek and Holub 2014). The gilding process was dangerous for the workers, because mercury vapours were released by the heating of gold and mercury alloys (Ullén et al. 2021). Those practices would lead to mercury intoxication during life, producing high levels of mercury in human bones – however, once manufactured those objects were used by people, a practice that could have increased the mercury exposition in life without being extremely hazardous for the person. The concentrations found in Tyresta for the human bones (1.55–63.29 ng∙g^− 1) are similar to those found in control or low-exposure populations (see Álvarez-Fernández et al. 2020). It is difficult to know the amount of mercury released during cremation, but if it was not significant, the observed values would be in line with moderate-to-low environmental metal pollution as the ones observed in post-Roman Lanzada (López-Costas et al. 2020). However, since mercury is volatile at high temperatures, we hypothesize that the post-mortem concentration before cremation could be higher. At least, despite being cremated, it is higher in human bones than in the unburnt animal bones, suggesting other source rather than the environmental that was described as the primary one in A Lanzada.

We have shown here that changes in colour due to cremation are highly correlated to changes in bone spectroscopic properties, indicating that the external changes and the physicochemical changes are highly intertwined. This opens the door for the use of analytical colour methodologies for the estimation of chemical changes in bone.

The macroscopic differences for the two categories of calcined bones (m-Stiner 6A and 6B) have been demonstrated here to correspond to two different burning categories and not a result of diagenetical alterations, at least for the samples analysed in the present study.

Direct mercury analyses have shown that Tyresta human samples contain mercury even after being heated at high temperatures. The concentration of the human bones is higher than the animals, even though they all seem to have lived in the same period, discarding an environmental source. Since mercury gets volatilized with temperature, it is possible that ante-mortem concentration in the human individual was higher, and we hypothesize that it would be connected to cultural practices (e.g. Contact with mercury gilded objects). The study of cremated bones by FTIR-ATR (relevant absorbances and MIR-indices) combined with chemometric analyses, has demonstrated to provide valuable information on the changes that occur during cremation. In particular, the allocation of the 1088 cm^− 1 band in a different PC than the main ν₃PO₄³⁻ band allows us to propose a new index (_TPV) that is worth exploring in future investigations.

Regarding the archaeological implications, the study allowed to increase our knowledge on the Tyresta site, in which we have found that: i) bones were probably heated at temperatures lower than 1100°C but higher than 900°C; ii) it is likely that no human intervention was done once the fire was started; iii) the ‘Badger layer’ was probably intentional and the bones found in it probably belong to another cremation event or even came from a different burial; and iv) mercury exposure was presumable low during lifetime of the cremated human and non-human individuals, but it could be detected.

Competing interests

Financial interests: none Non-financial interests: none

Funding

This project was funded by Grupos de Referencia Competitiva (ED431C 2021/32) by Xunta de Galicia, and by Galician Government through the support to research centres of the Galician university system, co-funded by FEDER (UE) (ED431G/2023/12), and by the European Union (ERC Consolidator Grant, PollutedPast, 101087832). Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Council. Neither the European Union nor the granting authority can be held responsible for them. OLC is funded by Ramón y Cajal 2020 (RYC2020-030531-I) del Ministerio de Ciencia e Innovación. MCP is funded by the European Union (ERC Consolidator Grant, PollutedPast, 101087832).

Acknowledgments

We would like to thank Kerstin Lidén for providing access to the facilities in which part of the experimental part of the study was performed and for her help and support during the stay at Stockholm University. We would like to thank Aikaterini Glykou for her advice and guidance during the laboratory proceedings. Also, we would like to thank Sven Isaksson for his guidance in the particularities of the FTIR equipment used for this paper. The authors acknowledge to CRETUS, Centre for Cross-disciplinary research in Environmental Technologies for financial support. We also would like to thank Markus Frid and Emelie Jonsson for their help in the selection of the samples for analysis and their advice in the species identification.

Compliance with Ethical Standards

We followed the ethical standards proposed by Squires et al. in: Squires, K., Booth, T., and Roberts, C. A. (2019). The ethics of sampling human skeletal remains for destructive analyses. Ethical approaches to human remains: A global challenge in bioarchaeology and forensic anthropology, 265-297. https://doi.org/10.1007/978-3-030-32926-6_12

Autor contributions

Marta Colmenares Prado: Conceptualization, methodology, formal analysis, investigation, writing-original draft, visualization, funding acquisition. Antonio Martínez Cortizas: Conceptualization, methodology, formal analysis, investigation, resources, writing-original draft, supervision. Jan Storå: Conceptualization, methodology, resources, writing-original draft, visualization, supervision. Mattias Pettersson: investigation, writing-original draft, visualization. Olalla López Costas: Conceptualization, formal analysis, investigation, writing-original draft, supervision, funding acquisition.

Álvarez-Fernández N, Martínez Cortizas A (2020) Andurinha: make spectroscopic data processing easier. Rpackage version 0.0.2 https://CRAN.R-project.org/package_andurinha
Álvarez-Fernández N, Martínez Cortizas A, López-Costas O (2020) Atmospheric mercury pollution deciphered through archaeological bones. Journal of archaeological science 119:105159. https://doi.org/10.1016/j.jas.2020.105159
Álvarez-Fernández N, Martínez Cortizas A, López-Costas O (2022) Structural equation modelling of mercury intra-skeletal variability on archaeological human remains. Science of the Total Environment 851:158015. https://doi.org/10.1016/j.scitotenv.2022.158015
Awad S, Loubar K, Tazerout M (2014) Experimental investigation on the combustion, performance and pollutant emissions of biodiesel from animal fat residues on a direct injection diesel engine. Energy 69:826-836.
Berna F (2010) Bone alteration and diagenesis. In: Artioli, G (ed) Scientific Methods and Cultural Heritage: An Introduction to the Application of Materials Science to Archaeometry and Conservation Science. Oxford University Press, Oxford, pp 364–367
Bratt P (2008) Makt uttryckt i jord och sten: stora högar och maktstrukturer i Mälardalen under järnåldern. Dissertation, Stockholm university,
Carvalhinho Windmöller C, Silva NC, Morais Andrade PH, Aparecida Mendes L, Magalhães do Valle C (2017) Use of a direct mercury analyzer for mercury speciation in different matrices without sample preparation. Analytical Methods 9(14):2159-2167. https://doi.org/10.1039/C6AY03041F
Castorina F, Masi U, Giorgini E, Mori L, Tafuri MA, Notarstefano V (2023) Evidence for Mild Diagenesis in Archaeological Human Bones from the Fewet Necropolis (SW Libya): New Insights and Implications from ATR–FTIR Spectroscopy. Applied Sciences 13(2):687. https://doi.org/10.3390/app13020687
Colmenares-Prado M, Cortizas AM, Veiga-Rilo C, López-Costas O (2025) Application of FTIR spectroscopy to infer ante-and post-mortem changes in archaeological human bone. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 330:125675. https://doi.org/10.1016/j.saa.2024.125675
DeHaan JD (2012) Sustained combustion of bodies: some observations. Journal of Forensic Sciences 57(6):1578-1584. https://doi.org/10.1111/j.1556-4029.2012.02190.x
Demšar J, Curk T, Erjavec A, Gorup Č, Hočevar T, Milutinovič M, Možina M, Polajnar M, Toplak M, Starič A, Stajdohar M, Umek L, Žagar L, Žbontar J, Žitnik M, Zupan B (2013) Orange: data mining toolbox in Python. the Journal of machine Learning research 14(1):2349-2353.
Etok SE, Valsami-Jones E, Wess TJ, Hiller JC, Maxwell CA, Rogers KD, Manning DAC, White ML, Lopez-Capel E, Collins MJ, Buckley M, Penkman KEH, Woodgate SL (2007) Structural and chemical changes of thermally treated bone apatite. Journal of Materials Science 42:9807-9816. https://doi.org/10.1287/mnsc.2022.00599
Evans M, Faulkner P, Asmussen B (2022) Investigating intentionality of burning through macroscopic taphonomy in complex legacy funerary assemblages: Opportunities and challenges. Journal of Archaeological Science: Reports 41:103243. https://doi.org/10.1016/j.jasrep.2021.103243
Festa G, Andreani C, Baldoni M, Cipollari V, Martínez-Labarga C, Martini F, Rickards O, Rolfo MF, Sarti L, Volante N, Senesi R, Stasolla FR, Parker SF, Vassalo AR, Mamede AP, Batista de Carvalho LAE, Marques MPM (2019) First analysis of ancient burned human skeletal remains probed by neutron and optical vibrational spectroscopy. Science Advances 5(6):eaaw1292. https://doi.org/10.1126/sciadv.aaw1292
Figueiredo MM, Gamelas JAF, Martins AG (2012) Characterization of bone and bone-based graft materials using FTIR spectroscopy. Infrared spectroscopy-life and biomedical sciences: 315-338.
Fleet ME (2009) Infrared spectra of carbonate apatites: ν2-Region bands. Biomaterials 30(8):1473-1481. https://doi.org/10.1016/j.biomaterials.2008.12.007
French KM, Crowder C, Crabtree PJ (2022) Integrating histology in the analysis of multispecies cremations: A case study from early medieval England. International Journal of Osteoarchaeology 32(6):1253-1263. https://doi.org/10.1002/oa.3160
Galeano S, García‐Lorenzo ML (2014) Bone mineral change during experimental calcination: an X‐ray diffraction study. Journal of forensic sciences 59(6):1602-1606. https://doi.org/10.1111/1556-4029.12525
Gallo G, Fyhrie M, Paine C, Ushakov SV, Izuho M, Gunchinsuren B, Zwyns N, Navrotsky A (2021) Characterization of structural changes in modern and archaeological burnt bone: Implications for differential preservation bias. PloS one 16(7):e0254529. https://doi.org/10.1371/journal.pone.0254529
Gallo G, Ushakov SV, Navrotsky A, Stahlschmidt MC (2023). Impact of prolonged heating on the color and crystallinity of bone. Archaeological and Anthropological Sciences 15(9):143. https://doi.org/10.1007/s12520-023-01842-0
García F, Ortega A, Domingo JL, Corbella J (2001) Accumulation of metals in autopsy tissues of subjects living in Tarragona county, Spain. Journal of Environmental Science and Health, Part A 36(9):1767-1786. https://doi.org/10.1081/ESE-100106258
Geiger S B, Weiner S (1993) Fluoridated carbonatoapatite in the intermediate layer between glass ionomer and dentin. Dental Materials 9(1):33-36. https://doi.org/10.1016/0109-5641(93)90102-V
Gonçalves D, Rosa J, Brandão AL, Martins A, Neves C, Diniz M, Arnaud JM, Marques MPM, Batista de Carvalho LA (2023) Infrared Spectroscopy to Assess Manufacturing Procedures of Bone Artefacts from the Chalcolithic Settlement of Vila Nova de São Pedro (Portugal). Applied Sciences 13(14):8280. https://doi.org/10.3390/app13148280
Gonçalves D, Vassalo AR, Mamede AP, Makhoul C, Piga G, Cunha E, Marques MPM, Batista de Carvalho LA (2018) Crystal clear: Vibrational spectroscopy reveals intrabone, intraskeleton, and interskeleton variation in human bones. American Journal of Physical Anthropology 166(2):296-312. https://doi.org/10.1002/ajpa.23430
Greiner M, Rodríguez-Navarro A, Heinig MF, Mayer K, Kocsis B, Göhring A, Toncala A, Grupe G, Schmahl WW (2019) Bone incineration: An experimental study on mineral structure, colour and crystalline state. Journal of Archaeological Science: Reports 25:507-518. https://doi.org/10.1016/j.jasrep.2019.05.009
Grunenwald A, Keyser C, Sautereau AM, Crubézy E, Ludes B, Drouet C (2014) Revisiting carbonate quantification in apatite (bio) minerals: a validated FTIR methodology. Journal of archaeological science 49:134-141. https://doi.org/10.1016/j.jas.2014.05.004
Habelitz S, Pascual L, Durán A (2001) Transformation of tricalcium phosphate into apatite by ammonia treatment. Journal of Materials Science 36:4131-4135. https://doi.org/10.1023/A:1017996217452
Herrmann B (1988) Behandlung von Leichenbrand. Anthropologie. Handbuch der vergleichenden Biologie des Menschen 4:576-585.
Hiller JC, Thompson TJ, Evison MP, Chamberlain AT, Wess TJ (2003) Bone mineral change during experimental heating: an X-ray scattering investigation. Biomaterials 24(28):5091-5097. https://doi.org/10.1016/S0142-9612(03)00427-7
Holck P (1997) Cremated bones: a medical-anthropological study of an archaeological material on cremation burials. Dissertation, University of Oslo, Oslo.
Hüls CM, Erlenkeuser H, Nadeau MJ, Grootes PM, Andersen N (2010) Experimental study on the origin of cremated bone apatite carbon. Radiocarbon 52(2):587-599.https://doi.org/10.1017/S0033822200045628
Innocenti D, Simonit F, Desinan L (2024) Burning temperature and bone modification: The cremation dynamics. Legal Medicine 67:102376. https://doi.org/10.1016/j.legalmed.2023.102376
Iregren E (1997) Why animal bones in human graves: an attempt to interpret animals present in Iron age cremations in Sweden. In: Smits E, Iregren E, Drusini AG (eds) Cremation Studies in Archaeology. LOGOS Edizione, Padova, pp 9-31.
Ježek M, Holub M (2014) Touchstones and mercury at Hedeby. Praehistorische Zeitschrift 89(1):193-204. https://doi.org/10.1515/pz-2014-0009
Kjellström A, Gummesson S, Storå J, Sauvage R (2024) Cremation and Depositional Practices of Cremated Remains at Skeiet. In: Sauvage R (ed) From Life to Death in Iron Age and Medieval Vinjefjord. Cappelen Damm Akademisk, pp 219–238
Krap T, Leenstra A, Oostra RJ, Duijst W (2024) Temperature estimation accuracy based on colourimetry of embalmed human and fresh non-human burned bone. International Journal of Legal Medicine 138(5):2107-2111. https://doi.org/10.1007/s00414-024-03239-7
Krap T, Ruijter JM, Nota K, Karel J, Burgers AL, Aalders MC, Oostra RJ, Duijst W (2019) Colourimetric analysis of thermally altered human bone samples. Scientific Reports 9(1):8923. https://doi.org/10.1038/s41598-019-45420-8
Krap T, van de Goot FR, Oostra RJ, Duijst W, Waters-Rist AL (2017) Temperature estimations of heated bone: A questionnaire-based study of accuracy and precision of interpretation of bone colour by forensic and physical anthropologists. Legal Medicine 29:22-28. https://doi.org/10.1016/j.legalmed.2017.08.001
Kumar MS, Kerihuel A, Bellettre J, Tazerout M (2006) Ethanol animal fat emulsions as a diesel engine fuel–part 2: engine test analysis. Fuel 85(17-18):2646-2652. https://doi.org/10.1016/j.fuel.2006.05.023
Lebon M, Reiche I, Bahain JJ, Chadefaux C, Moigne AM, Fröhlich F, Sémah F, Schwarcz HP, Falguéres C (2010) New parameters for the characterization of diagenetic alterations and heat-induced changes of fossil bone mineral using Fourier transform infrared spectrometry. Journal of Archaeological Science 37(9):2265-2276. https://doi.org/10.1016/j.jas.2010.03.024
Lebon M, Reiche I, Gallet X, Bellot-Gurlet L, Zazzo A (2016) Rapid quantification of bone collagen content by ATR-FTIR spectroscopy. Radiocarbon 58(1):131-145. https://doi.org/10.1017/RDC.2015.11
Legan L, Leskovar T, Črešnar M, Cavalli F, Innocenti D, Ropret P (2020) Non-invasive reflection FTIR characterization of archaeological burnt bones: reference database and case studies. Journal of Cultural Heritage 41:13-26. https://doi.org/10.1017/RDC.2015.11
LeGeros JP, LeGeros RZ (1983) Carbonate analyses of synthetic, mineral and biological apatites American Association for Dental ResearchAbstract 62:808.
Leskovar T, Zupanič Pajnič I, Jerman I (2023) Dealing with minor differences in bone matrix: can spectra follow the DNA preservation?. Australian Journal of Forensic Sciences 55(2):203-222. https://doi.org/10.1080/00450618.2021.1948102
Leskovar T, Cavalli F, Legan L, Innocenti D, Ropret P, Črešnar M (2025) Integrating Attenuated Total Reflectance–Fourier Transform Infrared Spectroscopy and Multidetector Computed Tomography for Analysis of Heat-Induced Changes in Bone. Materials 18(4):742. https://doi.org/10.3390/ma18040742
López-Costas O, Kylander M, Mattielli N, Álvarez-Fernández N, Pérez-Rodríguez M, Mighall T, Bindler R, Martínez Cortizas A (2020) Human bones tell the story of atmospheric mercury and lead exposure at the edge of Roman World. Science of the Total Environment 710:136319. https://doi.org/10.1016/j.scitotenv.2019.136319
Mamede AP, Gonçalves D, Marques MPM, Batista de Carvalho LA (2018a) Burned bones tell their own stories: A review of methodological approaches to assess heat-induced diagenesis. Applied Spectroscopy Reviews 53(8):603-635. https://doi.org/10.1080/05704928.2017.1400442
Mamede AP, Vassalo AR, Piga G, Cunha E, Parker SF, Marques MPM, Batista de Carvalho LAE, Gonçalves D (2018b) Potential of bioapatite hydroxyls for research on archeological burned bone. Analytical Chemistry 90(19):11556-11563. https://doi.org/10.1021/acs.analchem.8b02868
Marques MPM, Mamede AP, Vassalo AR, Makhoul C, Cunha E, Gonçalves D, Parker SF, Batista de Carvalho LAE (2018) Heat-induced bone diagenesis probed by vibrational spectroscopy. Scientific Reports 8(1):15935. https://doi.org/10.1038/s41598-018-34376-w
Marques MP, Gonçalves D, Mamede AP, Coutinho T, Cunha E, Kockelmann W, Parker SF, Batista de Carvalho LAE (2021) Profiling of human burned bones: oxidising versus reducing conditions. Scientific Reports 11(1):1361. https://doi.org/10.1038/s41598-020-80462-3
Martinez Cortizas A, López-Costas O (2024) A ATR-FTIR Study of Extracted Bioapatite from Archaeological Bone. https://doi.org/10.5772/intechopen.1007941
Maurer AF, Zeitoun V, Bardin J, Millard AR, Ségalen L, Guérin F, Saliège JF, Person A (2023) Multifactorial approach to describe early diagenesis of bones: The case study of the Merovingian cemetery of Saint-Linaire (France). Quaternary International 660:42-55. https://doi.org/10.1016/j.quaint.2023.03.003
Mayyas AS, Khrisat BR, Hoffmann T, El Khalili MM (2017) Fuel for lamps: organic residues preserved in Iron age lamps excavated at the site of Sahab in Jordan. Archaeometry 59(5):934-948. https://doi.org/10.1111/arcm.12288
McKinley JI (2015) In the Heat of the Pyre. In: Schmidt CW, Symes SA (eds), The Analysis of Burned Human Remains, 2nd edn. Elsevier/Academic Press, Amsterdam, pp 163–184. https://doi.org/10.1016/C2013-0-18935-0
Michel V, Bocherens H, Théry‐Parisot I, Valoch K, Valensi P (2006) Coloring and preservation state of faunal remains from the neanderthal levels of Kůlna Cave, Czech republic. Geoarchaeology 21(5):479-501. https://doi.org/10.1002/gea.20116
Panagos P, Jiskra M, Borrelli P, Liakos L, Ballabio C (2021) Mercury in European topsoils: Anthropogenic sources, stocks and fluxes. Environmental Research 201:111556. https://doi.org/10.1016/j.envres.2021.111556
Person A, Bocherens H, Mariotti A, Renard M (1996) Diagenetic evolution and experimental heating of bone phosphate. Palaeogeography, Palaeoclimatology, Palaeoecology 126(1-2):135-149. https://doi.org/10.1016/S0031-0182(97)88906-7
Price N, Raffield B (2018) Vikings. Milton: Taylor & Francis Group. https://doi.org/10.4324/9780203483251
Querido W, Ailavajhala R, Padalkar M, Pleshko N (2018) Validated approaches for quantification of bone mineral crystallinity using transmission Fourier transform infrared (FT-IR), attenuated total reflection (ATR) FT-IR, and Raman spectroscopy. Applied spectroscopy 72(11):1581-1593. https://doi.org/10.1177/0003702818789165
Rasmussen KL, Boldsen JL, Kristensen HK, Skytte L, Hansen KL, Mølholm L, Grootes PM, Nadeau M-J, Flöche Eriksen KM (2008) Mercury levels in Danish Medieval human bones. Journal of archaeological science 35(8):2295-2306. https://doi.org/10.1016/j.jas.2008.03.003
Rasmussen KL, Skytte L, D'Imporzano P, Thomsen PO, Sovso M, Boldsen JL (2017) On the distribution of trace element concentrations in multiple bone elements in 10 Danish medieval and post-medieval individuals. American Journal of Physical Anthropology 162(1):90-102. https://doi.org/10.1002/ajpa.23099
Rasmussen KL, Skytte L, Jensen AJ, Boldsen JL (2015) Comparison of mercury and lead levels in the bones of rural and urban populations in Southern Denmark and Northern Germany during the Middle Ages. Journal of Archaeological Science: Reports 3:358-370. https://doi.org/10.1016/j.jasrep.2015.06.021
Reidsma FH, van Hoesel A, van Os BJ, Megens L, Braadbaart F (2016) Charred bone: Physical and chemical changes during laboratory simulated heating under reducing conditions and its relevance for the study of fire use in archaeology. Journal of Archaeological Science: Reports 10:282-292. https://doi.org/10.1016/j.jasrep.2016.10.001
Rubio L, Díaz-Vico R, Smith-Fernández I, Smith-Fernández A, Suárez J, Martin-de-Las-Heras S, Santos I (2020) Spectrophotometric color measurement to assess temperature of exposure in cortical and medullar heated human bones: A preliminary study. Diagnostics 10(11):979. https://doi.org/10.3390/diagnostics10110979
Shahack-Gross R, Bar-Yosef O, Weiner S (1997) Black-coloured bones in Hayonim Cave, Israel: differentiating between burning and oxide staining. Journal of archaeological Science 24(5):439-446 https://doi.org/10.1006/jasc.1996.0128
Shaw CL (2022) An evaluation of the infrared 630 cm− 1 OH libration band in bone mineral as evidence of fire in the archaeological record. Journal of Archaeological Science: Reports 46:103655. https://doi.org/10.1016/j.jasrep.2022.103655
Sigvallius B (1994) Funeral pyres: Iron Age cremations in North Spånga. Dissertation, Stockholm University.
Snoeck C, Brock F, Schulting RJ (2014b) Carbon exchanges between bone apatite and fuels during cremation: impact on radiocarbon dates. Radiocarbon 56(2):591-602. https://doi.org/10.2458/56.17454
Snoeck C, Lee-Thorp JA, Schulting RJ (2014a) From bone to ash: Compositional and structural changes in burned modern and archaeological bone. Palaeogeography, palaeoclimatology, palaeoecology 416:55-68. https://doi.org/10.1016/j.palaeo.2014.08.002
Sponheimer M, Lee-Thorp JA (1999) Alteration of enamel carbonate environments during fossilization. Journal of Archaeological Science 26(2):143-150. https://doi.org/10.1016/j.palaeo.2014.08.002
Stamataki E, Kontopoulos I, Salesse K, McMillan R, Veselka B, Sabaux C, Annaert R, Boudin M, Capuzzo G, Claeys P, Dalle S, Hlad M, Sengeløv A, Vercauteren M, Warmenbol E, Tys D, De Mulder G, Snoeck C (2021) Is it hot enough? A multi-proxy approach shows variations in cremation conditions during the Metal Ages in Belgium. Journal of Archaeological Science 136:105509. https://doi.org/10.1016/j.jas.2021.105509
Stathopoulou ET, Psycharis V, Chryssikos GD, Gionis V, Theodorou G (2008) Bone diagenesis: New data from infrared spectroscopy and X-ray diffraction. Palaeogeography, Palaeoclimatology, Palaeoecology 266(3-4):168-174. https://doi.org/10.1016/j.palaeo.2008.03.022
Stiner MC, Kuhn SL, Weiner S, Bar-Yosef O (1995) Differential burning, recrystallization, and fragmentation of archaeological bone. Journal of archaeological science 22(2):223-237. https://doi.org/10.1006/jasc.1995.0024
Symes SA, Rainwater CW, Chapman EN, Gipson DR, Piper AL (2015) Patterned thermal destruction of human remains in a forensic setting. In: Schmidt CW, Symes SA (eds), The Analysis of Burned. Human Remains, 2nd edn. Elsevier/Academic Press, Amsterdam, pp 17-59 https://doi.org/10.1016/B978-012372510-3.50004-6
Thompson TJU, Gauthier M, Islam M (2009) The application of a new method of Fourier Transform Infrared Spectroscopy to the analysis of burned bone. Journal of Archaeological Science 36(3):910-914. https://doi.org/10.1016/j.jas.2008.11.013
Thompson TJU, Islam M, Bonniere, M (2013) A new statistical approach for determining the crystallinity of heat-altered bone mineral from FTIR spectra. Journal of Archaeological Science 40(1):416-422. https://doi.org/10.1016/j.jas.2012.07.008
Thompson T. (2015) The Archaeology of Cremation: Burned human remains in funerary studies. Oxbow Books, Oxford and Philadelphia.
Thompson TJ, Gonçalves D, Squires K, Ulguim P (2017) Thermal alteration to the body. Taphonomy of Human Remains: Forensic Analysis of the Dead and the Depositional Environment: Forensic Analysis of the Dead and the Depositional Environment 318-334. https://doi.org/10.1002/9781118953358.ch21
Trueman CN, Behrensmeyer AK, Tuross N, Weiner S (2004) Mineralogical and compositional changes in bones exposed on soil surfaces in Amboseli National Park, Kenya: diagenetic mechanisms and the role of sediment pore fluids. Journal of Archaeological Science 31(6): 721-739. https://doi.org/10.1016/j.jas.2003.11.003
Trueman CN, Privat K, Field J (2008) Why do crystallinity values fail to predict the extent of diagenetic alteration of bone mineral?. Palaeogeography, Palaeoclimatology, Palaeoecology 266(3-4):160-167 https://doi.org/10.1016/j.palaeo.2008.03.038
Ullén I, Nord AG, Tronner K (2021) Metal Analysis of Gilded Brooches from Viking Age Birka in South-central Sweden. Archäologisches Korrespondenzblatt 51(1):135-147. https://doi.org/10.11588/ak.2021.1.86996
Végh EI, Márquez-Grant N, Schulting RJ (2022) Physicochemical changes in bone bioapatite during the late postmortem interval pre-and post-burning. Applied spectroscopy 76(9):1080-1099.
Wang Q, Li W, Liu R, Zhang K, Zhang H, Fan S, Wang Z (2019) Human and non-human bone identification using FTIR spectroscopy. International journal of legal medicine 133:269-276. https://doi.org/10.1007/s00414-018-1822-8
Weiner S, Bar-Yosef O (1990) States of preservation of bones from prehistoric sites in the Near East: a survey. Journal of Archaeological Science 17(2):187-196. https://doi.org/10.1016/0305-4403(90)90058-D
Zazzo A, Saliège J-F, Lebon M, Lepetz S, Moreau C (2012) Radiocarbon dating of calcined bones: insights from combustion experiments under natural conditions. Radiocarbon 54(3-4):855-866.https://doi.org/10.1017/S0033822200047500
Zazzo A, Lebon M, Chiotti L, Comby C, Delqué-Količ E, Nespoulet R, Reiche I (2013) Can we use calcined bones for 14C dating the Paleolithic?. Radiocarbon 55(3):1409-1421. https://doi.org/10.1017/S0033822200048347

No competing interests reported.

SupplementaryInformation.docx

Download PDF

Editorial decision: Revision requested
19 Dec, 2025
Reviews received at journal
14 Oct, 2025
Reviews received at journal
14 Oct, 2025
Reviews received at journal
13 Oct, 2025
Reviewers agreed at journal
26 Sep, 2025
Reviewers agreed at journal
25 Sep, 2025
Reviewers agreed at journal
24 Sep, 2025
Reviewers invited by journal
20 Sep, 2025
Editor assigned by journal
04 Aug, 2025
Submission checks completed at journal
04 Aug, 2025
First submitted to journal
01 Aug, 2025

You are reading this latest preprint version

Using FTIR-ATR, analytical colour and mercury for unravelling the cremation ritual of Tyresta Viking Age burial mound (South-Central Sweden)

Status:

Version 1

Abstract

Figures

1. Introduction

2. Material and methods

2.1. Tyresta Archaeological site

2.2. Material

2.3. Macroscopic Colour

2.4. Analytical colour

2.5. Mercury concentration

2.6 FTIR-ATR

2.6. Statistical analysis

3. Results

3.1. Macroscopic colour

3.1.2. Analytical colour

3.2. Total mercury (Hg)

3.3. FTIR-ATR spectra and indices

3.3.3. Mercury (Hg) and FTIR

3.4. MIR-indices

3.5. Principal component analysis (PCA, PCAN, PCAC) and Discriminant analysis (DA)

3.5. Spectroscopic differences between groups

4. Discussion

4.1. Mercury content in bones

4.2. Colour as temperature indicator

4.3. Interpreting the PCA results

4.4. Transformation of bone with temperature: shift between m-Stiner grades 3 and 4.

4.5. Transformation with temperature: formation of new bands: 2012, 1088, 961, 699 cm− 1

4.5.1. ν1,3 PO3 2− bands (1088, 961 cm− 1)

4.5.2. Cyanamide formation

4.6. Bone type and species

4.7. Variations with time of exposure

4.8. Archaeological implications in Tyresta site

5. Conclusions

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1

3.5. Principal component analysis (PCA, PCA_N, PCA_C) and Discriminant analysis (DA)

4.5. Transformation with temperature: formation of new bands: 2012, 1088, 961, 699 cm^− 1

4.5.1. ν_1,3 PO₃ ²⁻ bands (1088, 961 cm^− 1)