β-Tricalcium Phosphate Interferes with the Assessment of Crystallinity in Burned Skeletal Remains

Laboratory of Forensic Anthropology, Centre for Functional Ecology, Department of Life Sciences, University of Coimbra, Calçada Martim de Freitas, 3000-456 Coimbra, Portugal Unidade de I&D “Química-Física Molecular”, Department of Chemistry, University of Coimbra, 3004-535 Coimbra, Portugal Unitat de Antropologia Biologica, Departament de Biologia Animal, Biologia Vegetal i Ecologia, Universitat Autonoma de Barcelona, 08193 Bellaterra, Spain Department of Chemistry and Pharmacy, University of Sassari, Via Vienna 2, 07100 Sassari, Italy Archaeosciences Laboratory, Directorate General for Cultural Heritage and LARC/CIBIO/InBIO, Rua da Bica do Marquês 2, 1300-087 Lisboa, Portugal Research Centre for Anthropology and Health (CIAS), Department of Life Sciences, University of Coimbra, Calçada Martim Freitas, 3000-456 Coimbra, Portugal

For example, burned bones from forensic settings include those of fire victims resulting from, among others, vehicle accidents, mass disasters, and house fires.In addition to accidents, in homicides, the victim's body may be purposely burned by the perpetrator in an attempt to destroy it, thus obstructing the investigation.Indeed, the effect of high temperatures on the human body can undermine and drastically complicate the bioanthropologists' examination.Regardless of the context, one of the key factors for the correct interpretation of the remains and the reconstruction of the incidents leading to burning is the estimation of the maximum exposure temperature [8][9][10].Micro-and ultrastructural analyses on burned skeletal remains are crucial for obtaining a reliable estimation of maximum burning temperature [11].While macroscopic alterations (e.g., surface colour) can be used to deduce an approximate temperature range [12][13][14][15], the investigation of the micro-and ultrastructural alterations of skeletal hard tissue exposed to high temperatures has proven to be crucial to get a reliable estimation of maximum temperature [8,9,13,[16][17][18][19][20][21][22][23].The bone which has been thermally altered shows an increase in crystallinity, exhibiting larger crystals and lower lattice strains [8,9,13,[16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31].
The analysis of burned remains is a highly complex process, and with new technologies available, a better insight can be gained.X-ray diffraction (XRD) often combined with Fourier-transform infrared spectroscopy attenuated total reflectance (ATR-IR) techniques is actually widely used to obtain primary material information in forensic and archaeological fields, such as the accuracy of temperature determination and the study of crystallinity [10].
Recently improved FT-IR approaches and statistical methods for the comparability of CI results have been established [40].However, the CI does not characterize individual crystal features (e.g., size or morphology) and may fail to describe adequately the complexity and heterogeneity of heat-induced processes [41].In fact, when bioapatite is subjected to a strong thermal treatment, we can find also a multiphase condition for the resultant product due to the transformation of a part of HA to the β-tricalcium phosphate (β-TCP) and other phases detected in different percentages among bones and teeth [23,30,42,43].The presence of β-TCP as well as the presence of other mineralogical phases due to various taphonomic effects can strongly alter the calculation of the CI and other ratios (C/P and Am/P or Am/C) to a nonnegligible level [41].
That is why a multidisciplinary approach is always advisable, possibly with the combined use of various physicochemical techniques.
The aim of this work is to demonstrate that only the combined use of XRD and ATR-IR techniques can document the various reactive transformations of the apatite phase.This is done by analyzing three experimentally burned human skeletons as well as 12 teeth at 1050 °C for one hour of residence.

Materials and Methods
The human bone samples were taken from three different unidentified skeletons, previously inhumed at the Capuchos cemetery (Santarém, Portugal) for a minimum of three years and afterwards donated to the University of Coimbra.They have the same provenance as the unclaimed skeletons of the 21st century identified skeletal collection housed at the  Table 1: List of the XRD phase investigation of the three modern human individuals, according to the Rietveld analysis.The uncertainty for the phase amount is calculated on the basis of statistical assumptions on the residual behaviour but generally holds about 10% of its value when the fit is carried out correctly.The value for the mostly used agreement factor R wp (which is a test for the goodness of the fit) in the refinement stage is also reported in the latter column.The average crystallite size of the hydroxyapatite mineral phase (1 Å = 10 −10 m) and CI from ATR-IR analysis have also been reported.[45].

Individual code
The skeletal remains were cleaned and macerated.The most superficial region of the bones was discarded with a scalpel to avoid possible contaminated samples and only then bone powder sampling took place.Samples were then concealed in Eppendorf pellets until the XRD analysis was performed.The CC/NI/16 samples comprised the humerus, radius, ulna, femur, tibia, calcaneus, talus (see Figure 1), and ribs 9 and 10, all from the left side.Additionally, sampling of a thoracic vertebra was carried out for this skeleton.The CC/NI/17 samples comprised the right ulna, radius, tibia, and fibula as well as two thoracic and two lumbar vertebrae.
Finally, the CC/NI/18 samples were composed of the clavicle, humerus, radius, and ulna, all from the right side.Samples from one cervical, two thoracic, and one lumbar vertebra were also collected.
The 12 human molar teeth employed in this study were kindly made available from the Department of Animal Biology, Plant Biology and Ecology, Autonomous University of Barcelona (Spain).
The experimental burning of the bones from the unidentified skeletons was carried out in an electric muffle (Barracha K-3, three-phased 14A model).The bones were all subjected to gradually increasing heating from room temperature to 1050 °C, which took 240 min to achieve.The muffle was then allowed to cool down to room temperature.In total, 48 samples were collected for XRD and ATR-IR analyses.
The 12 molar teeth for this experiment were heat-treated with a heating rate of 20 °C/min at 1050 °C for 60 minutes in air using a NEY muffle furnace.0.5 g of each sample was ball milled in an agate jar for one minute using a SPEX Mixer/ Mill model 8000, to get enough powder for the XRD and ATR-IR analyses.

XRD Analysis.
A small fraction (190 mg) of powdered bones was deposited in a dedicated sample holder for XRD analysis with a circular cavity of 25 mm in diameter and 2 mm in depth.The XRD patterns were collected using the Bruker D2 PHASER instrument working at a power of 30 Kv and 10 Ma in the Bragg-Brentano vertical alignment with a Cu-Ka tube emission (λ = 1 5418 Å).The width of divergent and antiscatter slits was 1 mm (0.61 °).Primary and secondary axial Soller slits of 2.5 °were also mounted with a linear detector LYNXEYE with 5 °opening and a monochromatisation by Ni foil for the Kβ radiation.The powder patterns were collected in the angular range 9-140 °in 2θ with a step size of 0.05 °.The collection time of each pattern was pursued for 47 min.Digitized diagrams were subjected to the analysis by the HighScore® and Match® programs which are able to locate the peak position in the 2θ reciprocal scale.The succession of peaks is compared with data from literature based on a search-match algorithm able to recognize the phase composition.The raw data were further analysed using the Rietveld approach for quantitative evaluation of phases.
The Rietveld method [46,47] is based on an iterative best-fit strategy of experimental data.We have made use of the MAUD (Material Analysis Using Diffraction) program which simulates the pattern by incorporating the instrument function and convolving the crystallographic model based on the knowledge of chemical composition and space group with selected texture and microstructure models [48].The program permits a selection of variables for the least squares minimization such as lattice parameters of the unit cell, atomic positions, temperature factors, occupancy of the sites, an/isotropic size, and strain broadening.
The success of the procedure is generally evaluated throughout a combination of integrated agreement factors (R wp is the most considered) and distribution of residuals [47].

ATR Analysis.
FT-IR spectra were collected in ATR mode with a Bruker Alpha Platinum-ATR interferometer in terms of absorbance versus wavenumber ν in the range 370-4000 cm −1 , with a resolution of 4 cm −1 .Each spectrum was obtained by averaging 512 interferograms.The loose  According to Rietveld analysis, the amount of bioapatite is 68 wt% and 32 wt% belongs to β-tricalcium phosphate.The residuals are reported on the top of the graph for ease of presentation.
powder was dispersed inside a hole cavity of spheroidal shape with its surface aligned to the plate defining it.

Crystallinity Index.
The crystallinity index adopted here is the same as has been used in the majority of archaeological applications.The absorption bands at 605 and 565 cm −1 were used following baseline correction, and the heights of these absorptions peaks were summed and then divided by the height of the minimum between them [32].
In detail, portlandite was found in 13 specimens (weight fraction range from 2 to 12 wt%), β-TCP in 10 specimens (from 1 to 32 wt%), lime in 27 specimens (from 1 to 5 wt%), calcite in 7 specimens (from 1 to 4 wt%), and buchwaldite in 1 case (2 wt%).Only in 4 cases, bones have remained unaltered, bioapatite 100% (CC/NI/18 individual), apart from the microstructure features assessed from peak sharpening and the organic component, which is expected to be removed from the bone with the thermal treatment carried out.
The average crystallite size of the examined bioapatite varies from a lower value of 1540 Å (CC/NI/16_left rib 10 anterior proximal) to an upper value of 2837 Å (CC/NI/ 18_vertebra II) (mean = 1950 Å).The crystallinity index varies from a lower value of 6.10 (CC/NI/18_vertebra II) to an upper value of 4.22 (CC/NI/17_right fibula) (mean = 5.14).Journal of Spectroscopy An emblematic case is represented by the sample CC/NI/ 17_right fibula, in which the bioapatite after the heat treatment at 1050 °C has partially transformed into β-TCP.The Rietveld analysis is reported in Figure 2. The experiment (data points) was fitted satisfactorily (R wp = 6.6%) with the full line after including structure information from the mineral bioapatite (68.0 wt%) and β-TCP (32.0 wt%).
The teeth have a lower crystallinity compared to the bones; in fact, the average crystallite size varies from a lower value of 1375 Å (T1 tooth) to an upper value of 1900 Å (T11 tooth) (mean = 1702 Å).The crystallinity index varies from a lower value of 3.24 (T1 tooth) to an upper value of 4.89 (T12 tooth) (mean = 3.85).
Figure 3 represents an extreme case (T1 sample), in which the analysis of the correspondent XRD pattern has established the presence of the 44% β-TCP phase for such specimen.
Figure 4 shows the ATR-IR spectra of three burned teeth (from bottom to top: T1, T3, and T4, resp.).The spectra are reported in the wavenumber ν ranging from 400 to 1500 cm −1 .
It is possible to recognize two main groups of bands in the range of 500-700 cm −1 and 1000-1200 cm −1 , which are generally assigned to the energy mode ν 4 of phosphate groups and ν 3 of phosphate groups, respectively.An additional peak at about 1123 cm −1 (see Figure 3) in T1 and T3 samples as indicated by arrows is attributable to β-TCP.
Particularly, the ν 4 band of the phosphates present in the ATR-IR spectrum can provide a variety of supporting information to XRD analysis, due to numerous deformations and displacements of the band shape and to the CI calculation.
In detail, Figure 5 highlights the band structure of ν 4 phosphate groups in the range 500-700 cm −1 of representative samples; Figure 5(a) shows the conventional pattern of CC/NI/18_vertebra II sample in which the bioapatite phase remained unchanged after the experimental heating (HA = 100%), as far as its crystal structure is concerned; the presence and the intensity of the shoulder at about 629 cm −1 indicate the occurrence of high thermal treatments [49].The larger average crystallite size (2837 Å) detected by XRD is coherent with the high value of CI (6.10).
Figure 5(b) documents the case (T4 tooth) in which the presence of β-TCP is rather substantial; in fact, the ν 4 band begins to slightly overlap with the ν 4 band of bioapatite (evidenced by the additional peaks at 456 and 555 cm −1 indicated by arrows).
Figure 5(c) represents the extreme case of T1 tooth when the presence of β-TCP is massive (44%); in fact, the ν 4 band appears strongly deformed due to the complete overlap with the ν 4 band of bioapatite (see the deformation of the peak at 555 cm −1 with the presence of two further peaks at 546 and 563 cm −1 , resp.).In such cases, the CI calculation is problematic and the value is unusually low when compared to that of other burned bone samples.

Discussion and Conclusive Remarks
The occurrence of β-TCP can follow from a chemical reaction according to which Ca 10 (PO 4 ) 6 (OH) 2 is subjected at high temperature, giving 3(Ca  °C that are particularly noteworthy because they present a significant amount of β-TCP (from bottom to top, T1 = 44 wt%, T3 = 42 wt%, and T4 = 25 wt%, resp.)detected with XRD.The deformation of the ν 4 band is observable, as well as the presence of an additional peak at 1123 cm −1 (in T1 and T2 specimens) due to the presence of β-TCP.

Journal of Spectroscopy
Recent studies [50] have attempted to clarify the transformation of a Ca-deficient synthetic apatite to β-TCP.Upon heating (calcining) to 710 ± 740 °C, the Ca-deficient apatite will transform to the low-temperature polymorph of β-TCP, with the loss of water as described by where CaO is missing with respect to the 2(Ca 5 (PO 4 ) 3 (OH)) conventional formula recognized for bioapatite.NaCaPO 4 and MgO are also observed in very small quantities probably as a consequence of impurities in the starting bone material.
The presence of β-TCP phase from bones appears to be sporadic and seems to occur only at high temperatures, around 1100 °C [43].Conversely, in our previous work, we observed a more systematic occurrence of β-TCP in teeth specimens at temperatures as low as 750 °C [41].
The reason why β-TCP appears in teeth at a relatively moderate temperature in comparison to bones is still obscure; it may be related to the fact that bones and teeth have different compositions and histologies, and further studies need to be addressed by acquiring information about chemical species and following the crystal structure parameters.
Our results demonstrate that the use of the CI for bioanthropological inferences such as those related to temperature estimation [8,9,35,39] and to the determination of bone quality and preservation [32,33] is not a straightforward procedure.Since the generation of β-TCP can affect the phosphate peaks located at the wavelength of interest for CI calculation, one should be especially careful whenever such peaks present anomalous shapes.tooth, and (c) T1 tooth with the corresponding XRD Rietveld phase evaluation.The distortions of ATR-IR ν 4 phosphate band with respect to the conventional expected behavior are strictly related to the amount of the β-TCP phase which was stimulated by the high-temperature treatment.Particularly, when the presence of β-TCP is massive (≥20%), the ν 4 band appears strongly deformed due to the overlap with the ν 4 band of bioapatite (see the deformation of the peak at 563 cm −1 ), and also, the CI calculation is problematic.8 Journal of Spectroscopy

Figure 1 :
Figure 1: Four examples of burned bones used in this study: (a) CC/NI/16, left calcaneus; (b) CC/NI/17, left talus; (c) CC/NI/16, comparison between unburned right femur and burned left femur; and (d) CC/NI/18, comparison between unburned and burned thoracic vertebra.The burned vertebra was affected by heat-induced shrinkage, thus explaining its substantial smaller size in comparison to the unburned vertebra.

Figure 2 :
Figure 2: CC/NI/17_right fibula represents one of the rare bone specimens showing a partial transformation to β-TCP.According to Rietveld analysis, the amount of bioapatite is 68 wt% and 32 wt% belongs to β-tricalcium phosphate.The residuals are reported on the top of the graph for ease of presentation.

Figure 3 :Figure 4 :
Figure3: The XRD pattern of T1 tooth heated up to 1050 °C for 1 hour.The phase analysis suggests that the material is a mixture of bioapatite (55 wt%) and β-TCP (Ca 3 (PO 4 ) 2 ) (44.0 wt%), with a minor contribution of MgO.

Figure 5 :
Figure 5: A detailed comparison of different FT-IR curves collected for three different bioapatite specimens: (a) CC/NI/18_vertebra II, (b) T4tooth, and (c) T1 tooth with the corresponding XRD Rietveld phase evaluation.The distortions of ATR-IR ν 4 phosphate band with respect to the conventional expected behavior are strictly related to the amount of the β-TCP phase which was stimulated by the high-temperature treatment.Particularly, when the presence of β-TCP is massive (≥20%), the ν 4 band appears strongly deformed due to the overlap with the ν 4 band of bioapatite (see the deformation of the peak at 563 cm −1 ), and also, the CI calculation is problematic.

Table 2 :
List of structure and microstructure parameters of bioapatite structure in teeth from phase analysis by XRD and the associated ATR-IR CI.
If water (H 2 O) evaporates completely, we observe the presence of CaO; alternatively, incomplete evaporation develops into Ca(OH) 2 .It depends on the speed of cooling after the burning process.