A universal curve of apatite crystallinity for the assessment of bone integrity and preservation

Abstract

The reliable determination of bioapatite crystallinity is of great practical interest, as a proxy to the physico-chemical and microstructural properties, and ultimately, to the integrity of bone materials. Bioapatite crystallinity is used to diagnose pathologies in modern calcified tissues as well as to assess the preservation state of fossil bones. To date, infrared spectroscopy is one of the most applied techniques for bone characterisation and the derived infrared splitting factor (IRSF) has been widely used to practically assess bioapatite crystallinity. Here we thoroughly discuss and revise the use of the IRSF parameter and its meaning as a crystallinity indicator, based on extensive measurements of fresh and fossil bones, virtually covering the known range of crystallinity degree of bioapatite. A novel way to calculate and use the infrared peak width as a suitable measurement of true apatite crystallinity is proposed, and validated by combined measurement of the same samples through X-ray diffraction. The non-linear correlation between the infrared peak width and the derived ISRF is explained. As shown, the infrared peak width at 604 cm^-1 can be effectively used to assess both the average crystallite size and structural carbonate content of bioapatite, thus establishing a universal calibration curve of practical use.

Figure 1

FTIR spectra of (a) fresh bone and (b) diagenetically altered archaeological bone. Baselines, the full width at half maximum (FWHM) of the phosphate peak at 1035 cm⁻¹, the width at 85% of height of the peak at 604 cm⁻¹ (FW85%), the width at 80% of height of the peak at 565 cm⁻¹ (FW80%) and the splitting factor (IRSF) are highlighted. Major vibrational bands attribution is also reported (see the Supplementary Information for a detailed description).

Figure 2

Correlation between the IRSF and FW85% among bioapatite samples and plot of the regression function.

Figure 3

Diffractogram of an archaeological bone sample from Sudan. The calculated pattern and the residual resulting from the Rietveld refinement (Rwp = 4.45%) is also displayed. The top-right panel shows the average shape of bioapatite crystallites, elongated along the c-axis, reconstructed from the spherical harmonics coefficients refined by Rietveld method.

Figure 4

(a) Simplified model of the ν₄(PO₄) vibrational mode representing the IR spectra (blue) resulting from the contribution of the 604 and 565 cm⁻¹ peaks, approximated by Gaussian curves; in light-blue is represented the change of the spectrum shape by varying the width of the two Gaussian curves. (b) Plot of the calculated IRSF against the width at 85% of the height of the modelled peaks; an exponential regression function fits the calculated points (R² = 0.996) and its first derivative shows the curve slope; the light blue area shows the variation range of IRSF and FW85% considering a possible shift of the two peaks centred position, resulting in a variation of the measured distance between the two maxima in the range 37–43 cm⁻¹ (according to the experimental dataset).

Figure 5

Plot showing the correlation between (a) the width at 80% of the height of the 565 cm⁻¹ peak (FW80%) and the width at 85% of the height of the 604 cm⁻¹ peak (FW85%); (b) the FWHM of the 1035 cm⁻¹ peak and the width at 85% of the height of the 604 cm⁻¹ peak (FW85%); (c) the variation of the 1020, 1090 and 1110 cm⁻¹ bands normalised to the 1035 cm⁻¹ band among all bioapatite samples. The relative standard deviation for these parameters is within 5%.

Figure 6

Plot showing (a) the non-linear correlation between the FW85% (FTIR spectroscopy) and the crystallite size along the c axis (XRPD); a tentative exponential fit is also displayed. The approximated linear correlation between the two parameters for archaeological bones from Sudan (R² = 0.92) is reported; (b) the linear correlation between IRSF (FTIR spectroscopy) and the crystallite size along the c axis (XRPD) considering the entire set of samples (R² = 0.90) and a subset of samples constituted by archaeological bones from Sudan (R² = 0.84); (c) Plot showing the correlation between the FW85% and the structural carbonate content of bioapatite; CO₃/PO₄ parameter was calculated for fresh bones and for a subset of archaeological bones not contaminated by secondary calcite. (d) Kernel density estimation of the FW85% parameter measured for all samples. Error bars show the relative standard deviation estimated for FTIR parameters (5%) and crystallite size (3%).