Visualizing Different Crystalline States during the Infrared Imaging of Calcium Phosphates

Abstract

Methods utilizing relatively simple mathematical operations during physical analyses to enable the visualization of otherwise invisible correlations and effects are of particular appeal to researchers and students in pedagogical settings. At the same time, discerning the amorphous phase from its crystalline counterpart in materials is challenging with the use of vibrational spectroscopy and is nowhere as straightforward as in phase composition analytical methods such as X-ray diffraction. A method is demonstrated for the use of first- and second-order differentiation of Fourier transform infrared spectra of calcium phosphates to distinguish their amorphous states from the crystalline ones based on the exact line positioning rather than on comparatively vaguer band broadening and splitting effects. The study utilizes a kinetic approach, focusing on the comparison of spectral features of amorphous precursors annealed in air at different temperatures and aged for different periods of time in an aqueous solution until transforming to one or a mixture of crystalline phases, including hydroxyapatite and α- and β-tricalcium phosphate. One of the findings challenges the concept of the nucleation lag time preceding the crystallization from amorphous precursors as a “dead” period and derives a finite degree of constructive changes occurring at the atomic scale in its course. The differential method for highlighting spectral differences depending on the sample crystallinity allows for monitoring in situ the process of conversion of the amorphous calcium phosphate phase to its crystalline analogue(s). One such method can be of practical significance for synthetic solid state chemists testing for the chemical stability and/or concentration of the reactive amorphous phase in these materials, but also for biologists measuring the maturity of bone and medical researchers evaluating its phase composition and, thus, the state of metabolic and mechanical stability.

Introduction

Transitions between amorphous and crystalline forms of the bone mineral accompany some of the most fundamental solid state reactions happening in our bodies[1]. Biomineralization, bone regeneration and bone remodeling all depend on the finely tuned transitions between these two phases of calcium phosphate (CP). Moreover, given the structural support that bone provides to the organism, this reversible phase transition can be thought of indirectly affecting the complete metabolism and physiological integrity of the human body.

The amorphous solid phase was detected in copious amounts in all the newly formed hard tissues in the human body, including bone[2], dentin[3], cementum[4] and enamel[5]. It is also known to be present in finite amounts in mature bones too[6], where it may reversibly transform to and from the crystalline phase in response to the metabolic demands of the organism. It coats virtually all the crystal faces of the bone mineral particles and is particularly prominent at their crystallization front[7]. The amorphous phase is also known to be a precursor for bone formation in a number of species other than humans[8,9,10], explaining its superior remineralization properties compared to crystalline CPs[11]. Simultaneously, biominerals other than CP, including ferrihydrates[12] and calcium carbonates[13], have been shown to form through the presence of transient amorphous phases. In a broader context, the versatility of physical forms adoptable by biominerals across all taxonomies, along with the fact that their formation proceeds strictly through amorphous intermediates, implies that a deeper understanding of the aforementioned transition can be a gateway to the methods for the control and synthesis of innumerable advanced biomimetic materials. In other words, it is conceivable that the most structurally sophisticated architectures on the nanoscale will be creatable strictly via utilizing non-classical mechanistic pathways that involve the use of amorphous building blocks. Devising the new and improving the existing analytical methods for distinguishing the amorphous from the crystalline states in biominerals can thus be of an immense importance for aspirations to achieve these higher synthetic goals.

The most common method for assessing the crystallinity of the bone mineral is X-ray diffraction (XRD). The downside of this method comes from its lack of sensitivity for discerning fine changes in the short-range order of amorphous or poorly crystalline domains, making it virtually impossible to draw a fine line between the two[14] without the use of not easily accessible synchrotron radiation addenda, such as the rapid acquisition pair distribution function technique[15]. Theoretically, however, in contrast to XRD, the rather inexpensive and more easily accessible vibrational spectroscopic methods are more sensitive to the short-range order surrounding the active atomic groups, the only problem being the comparatively low peak resolution of the resulting spectra. As a result, the most commonly used method for assessing the crystalline state in bone mineral has relied on the comparison of the full width at half-maxima (FWHM) of the dominant bands[16]. More seldom, comparisons of prominent peak intensity ratios have been used to analyze the crystallinity of CPs[17]. The delineation of exact spectral lines corresponding to each of the two phases, amorphous and crystalline, has not been possible based on the quantitative characterization of raw spectral features.

In this study, a method is described for the delineation of more exact spectral lines indicative of the amorphous and/or crystalline states in CPs in Fourier transform infrared (FTIR) spectroscopic analyses through the use of precise mathematical manipulation of the as-obtained spectra. The method is based on the relatively simple, first- and second-order differentiation of the raw spectra, a technique used for decades in FTIR spectral analyses of other materials[18,19,20,21]. The derivative FTIR spectroscopy allows for more specific and less ambiguous identification of fine and unresolved absorption peaks[22,23,24] than the spectral analyses based on deconvolution. Different CP phases have been resolved before in the raw FTIR spectra, but almost invariably by deconvolution of overlapping peaks[25,26,27], a method that implies a large degree of variability compared to the methods based on routine derivatizations of the spectra. One such method is used here to differentiate between amorphous and crystalline modifications of different CPs, including amorphous calcium phosphate (ACP), hydroxyapatite (HAp) and tricalcium phosphate (TCP). The main task for this study is, therefore, to prove its suitability for the distinction of different crystallization states in CPs as they coexist and coevolve. Logically, with the first application of this model for the analysis of the conversion of metastable ACP to stable crystalline polymorphs, meaningful insights about this process are deduced and commented on in the discussion. The given method is presented as a standalone model, but it is also placed in the context of its use in instruction in graduate and undergraduate courses on vibrational spectroscopy.

Materials and methods

ACP powder was synthesized by abruptly adding an aqueous solution containing 100 ml 0.5 M calcium nitrate (Ca(NO₃)₂, Fisher Scientific) and 7 ml 28 % ammonia (NH₄OH, Sigma Aldrich) into an aqueous solution comprising 100 ml 0.2 M monoammonium phosphate (NH₄H₂PO₄, Fisher Scientific) and 4 ml 28 % NH₄OH. The fine precipitate formed upon mixing was aged for 15 s, before it was collected, centrifuged (5 minutes at 3500 rpm), washed with deionized water, centrifuged again, washed with ethanol, and then lyophilized at temperatures ranging from −10 to −60°C and pressures ranging from 0.37 mbar to 0.1 mbar for 1 to 8 h. A portion of the powder was stored at 4 °C to prevent spontaneous transformation to HAp, while another portion was stored under ambient conditions for a month.

Portions of the samples (~ 10 mg) were annealed in a thermal analyzer (SETSYS 2400 CS Evolution SETARAM) to 100, 200, 400, 600, 800 and 1000 ^°C in air at the rate of 10 °C/min and cooled down at the rate of 20 °C/min. XRD analysis on fresh and annealed powders was carried out on an Ultima IV Rigaku diffractometer, in the 2θ range between 8° and 70°, with the step size of 0.02° and the scan rate of 5°/min. Diffractometric phase composition analysis was performed by comparing the experimentally obtained reflections with the standard patterns for HAp (ICDD: #09–432), α-TCP (ICDD #09–348) and β-TCP (ICDD #09–169). To estimate the average crystallite size (d_x), Debye-Scherrer’s equation was applied on the integral breadths of the most intense reflections in the diffractograms of ACP (2θ = 30.3 °), HAp (2θ = 31.0°), β-TCP (2θ = 31.8°) and α-TCP (2θ = 30.7 °) expressed in radians (β_1/2), using 1.5418 as the wavelength of Cu_Kα as the radiation source (λ):

$${\text{d}}_{\text{x}}=0.94\text{λ}/{\text{β}}_{1/2}\text{cosθ}$$ (Eq.1)

FTIR measurements were performed on a Thermo Scientific™ Nicolet™ iS™10 FTIR Spectrometer equipped with attenuated total reflectance (ATR) accessory in the 400 – 4000 cm⁻¹ spectral range. Raw spectra comprised 7468 data points, with the step size of 0.483 cm^-1. They were differentiated using a manual differentiation routine implemented on the data point columns comprising the spectra (Origin Pro 2018). First- and second-order differentiated spectra were smoothened using the manual Lowess routine and the proportion for span equaling 0.01 (Origin Pro 2018). Fig. 1 displays the FTIR spectra before and after the differentiation and smoothening routines, including the critical screenshots where the exact operations in the program are selected. Differentiation procedure follows a simple, default mathematical routine and its advantage for an instructional setting is also that it could be performed in free-access programs such as Spectragryph, Protea Spectrum Viewer or Microsoft Excel.

Fig.1.

The derivative method for the spectral analysis involves 1^st or 2^nd order differentiation of raw FTIR spectra (a) and the consequent smoothening of the differentiated spectra (b). The screenshots represent the computer interface for executing these two operations successively in Origin Pro 2018. Relevant regions of the CP spectra elaborated in the discussion are marked by the dashed red rectangles.

Results and discussion

Immediately removed from the parent solution, washed, dried and stored at temperatures lower than 4 °C, ACP preserves its amorphous structure, but during annealing it transforms to a mixture of α- and β-TCP at temperatures equal to or higher than 800 °C, as evidenced from XRD patterns presented in Fig. 2a. Also, as suggested by another set of XRD measurements, if aged for periods of time equal to or longer than 3 h in its parent aqueous solution, ACP transforms to HAp (Fig. 2b). CPs crystallized under both ambient conditions and elevated temperatures retained the nanocrystalline nature of all their phase constituents. Average crystallite sizes were under 35 nm for HAp formed both by ripening ACP at room temperature and by calcining ACP as well as for both TCP polymorphs formed by thermal annealing (Table 1).

Fig.2.

XRD patterns of ACP powders before (25 °C) and after annealing at 10 °C/min to different temperatures in the 25 – 800 °C range (a). XRD patterns of ACP aged after precipitation for different amounts of time in its parent solution at room temperature (b). Peaks indexed with ▪ denote reflections originating from HAp (ICDD: #09–432), whereas those indexed with o and * denote reflections originating from α-TCP (ICDD #09–348) and β-TCP (ICDD #09–169), respectively.

Table 1.

Average crystallite sizes (d_x) determined from Debye-Scherrer’s equation for different CP phases formed either by ripening (T = 21 °C) or by thermal annealing (T = 800 °C).

CP phase	dx (nm)	T (°C)
ACP	0.5	21
HAp	11.0	21
HAp	34.9	800
α-TCP	30.8	800
β-TCP	29.3	800

Fig. 3a shows the corresponding raw FTIR spectra in the 1200 – 400 cm⁻¹ range for ACP powders subjected to either no thermal treatment (25 °C) or annealed at different temperatures, ranging from 100 °C to 800 °C. Unlike the clear difference between the diffractograms of the amorphous phase (25 – 600 °C, 5 – 120 min) and the diffractograms of the crystalline phases (TCP @ 800 °C and HAp @ 180 min), the difference between the respective FTIR spectra is comparatively minor. As seen in Fig. 3a, all IR spectra of CP, irrespective of the annealing temperature, display the four characteristic phosphate vibration modes[28,29,30], including the non-degenerated symmetric stretch, v₁(P-O), the triply degenerated antisymmetric stretch, v₃(PO), the symmetric out-of-plane bend, v₂(O-P-O), and the triply degenerated antisymmetric bend, v₄(O-P-O). For non-annealed ACP, the v₁(P-O) band is centered at 953.2 cm⁻¹; the v₃(P-O) band at 1004.3 cm⁻¹; the v₂(O-P-O) band at 435.4 cm⁻¹; and the v₄(O-P-O) band at 549.2 cm^-1. At the onset of crystallization, between 600 and 800 °C, the spectral features slightly change, most notably displaying the splitting of the v₄(O-P-O) band into a doublet and the narrowing of the v₃(PO) band. While the changes occurring to the v₄ bending mode are evident from the raw spectra, the effect on the v₃ stretching mode is rather indistinct. The performance of the first-order differentiation, however, gives a distinct v₃ maxima separation between the amorphous and crystalline phases, amounting to 1002.5 cm⁻¹ – 977.3 cm⁻¹ = 25.2 cm⁻¹ (Fig. 3b). The derivative of the v₄ doublet of the crystalline phase is also a distinct new peak, overlapping with the troughs from by the amorphous phases (Fig. 3b). These spectral differences get further intensified by increasing the differentiation order from the first to the second, at which point the difference between the prominent v₃ negative peak components remains constant at 1016.7 cm⁻¹ – 991.5 cm⁻¹ = 25.2 cm⁻¹, while a new, positive v₃ peak component appears at 998.1 cm⁻¹, coinciding with the troughs yielded by the amorphous phases and acting as a key spectral feature applicable in distinguishing crystalline TCP from ACP using FTIR spectroscopy (Fig. 3c). Meanwhile, a negative v₄ peak component with the minimum at 604.8 cm⁻¹ becomes even more prominent than in the first-order derivatives, acting as yet another distinct feature of the crystalline phase. Here it should be noted that while the first derivatives peak at the highest slope (i.e., growth rate) points, the second derivatives peak at the inflection points whereat the growth rate changes sign and are thus even more sensitive to shoulders and other features of the fine structure of the spectral lines. As a result, all the spectral differences between ACP and TCP detectable in the first-order derivatives become magnified in their second-order analogues. Higher differentiation orders, albeit theoretically attemptable and practically implementable up to the fourth, successively reduce signal-to-noise ratio and diminish the reproducibility of spectral differences[31], for which reason they are rarely applied as spectral processing tools.

Fig.3.

FTIR spectra of ACP powders before (25 °C) and after annealing at 10 °C/min to different temperatures in the 25 – 800 °C range (a). First-order (b) and second-order (c) derivatives of the same FTIR spectra as in (a). Dashed rectangle in (b) denotes a region where the spectral features of the crystalline phase (800 °C) significantly deviate from its amorphous counterpart (25 – 600 °C).

Another interesting feature of the spectra as a function of the annealing temperature is that the absorption of the characteristic vibration modes intensifies with the temperature, indicating that whatever the short-range symmetry is adopted by ACP, it increases in direct proportion with the annealing temperature (Fig. 3a). One exception to this trend comes from a minor decrease in intensity from 25 °C to 100 °C, after which it begins to continually increase with temperature in the 100 – 600 °C range, conforming to an earlier report that the limited order present in the amorphous phase as it begins to transform to a crystalline phase must first be destroyed before a more lasting order gets reestablished[32].

Fig. 4a shows the raw FTIR spectra in the 1200 – 400 cm⁻¹ range for ACP and HAp prior to any thermal treatment. A direct comparison between these two CP phases is important because mineralization of bone proceeds from ionic precursors to ACP to poorly crystalline, nonstoichiometric HAp. In HAp, the non-degenerated symmetric stretch, v₁(P-O) is centered at 962.3 cm⁻¹; the triply degenerated antisymmetric stretch, v₃(P-O) is centered at 1021.2 cm⁻¹; the out-of-plane symmetric bend, v₂(O-P-O) is centered at 559.8 cm⁻¹; and the triply degenerated antisymmetric bend, v₄(O-P-O) is centered at 472.5 cm^-1. Expectedly, because of the lesser translational symmetry of the order around the active groups, these four bands become broadened in ACP and either lose their multiplet structure, as for v₄(O-P-O), or blend into the shoulder of a more intense nearby peak, as for v₂(O-P-O). The differentiation routine, as in the case of ACP → TCP transition (Fig. 3), here also magnifies the spectral differences between ACP and HAp and eventually allows for their unambiguous distinction. Because of the peak splitting effects, the derivative spectra of HAp exhibit peaks, both positive and negative, which are not present in ACP. However, even if these rough splitting effects were to be ignored, the differentiation allows for HAp to be distinguished from ACP based on a direct upshift of the most intense phosphate vibration modes of the former phase. Specifically, in the first-order derivatives (Fig. 4b), while the v₃(P-O) mode in the spectrum of HAp peaks at a frequency higher by 1006.8 cm⁻¹ – 981.7 cm⁻¹ = 25.1 cm⁻¹ than the v₃(P-O) mode of ACP, the v₄(O-P-O) mode peaks at a frequency higher by 547.9 cm⁻¹ – 531.5 cm⁻¹ = 16.4 cm^-1. In the second-order derivatives (Fig. 4c), the positive component of the v₃(P-O) mode in the spectrum of HAp peaks at a frequency higher by 993.7 cm¹ = 968.5 cm⁻¹ = 25.2 cm⁻¹ than the same component of the v₃(P-O) mode of ACP; for the negative component of the same mode, this difference amounts to 1020.0 cm⁻¹ – 995.9 cm⁻¹ = 24.1 cm^-1. At the same time, the positive component of the v₄(O-P-O) mode peaks at a frequency higher by 543.5 cm⁻¹ – 522.6 cm⁻¹ = 20.9 cm⁻¹ than the same component of the v₄(O-P-O) mode of ACP; for the negative component of the same mode, this difference equals 562.1 cm⁻¹ – 546.8 cm⁻¹ = 15.3 cm^-1.

Fig.4.

FTIR spectra of ACP powders aged after precipitation for different amounts of time in their parent solutions at room temperature (a). First-order (b) and second-order (c) derivatives of the same FTIR spectra as in (a).

Therefore, although a comparison between the centering of the two most intense phosphate bands in the FTIR spectra of HAp and ACP does not allow for the discernment of one against the other, both the first- and the second-order derivatives allow for this qualitative discernment to be unambiguously made. This distinction is less ambiguous and more straightforward to execute mathematically compared to the methods based on the deconvolution of P-O vibration modes in HAp in search of the downshift shoulders that originate from the amorphous component[33]. Specifically, while the differentiation of raw, unprocessed spectra is an unambiguous and readily translatable procedure, deconvolution implies variable results due to a number of different algorithms that are possible to use and that do not always converge, alongside the approximations inherent in each one of them. Fourier self-deconvolution (FSD) similarly requires an approximated input on the line shape width and the smoothing factor to reduce the bandwidth and resolve the overlapping spectral lines. Phenomenologically, the FSD procedure of transforming the spectrum back to an interferogram, then correcting for the exponential decay in its Lorentzian cosine waves before recreating the original spectrum with narrowed bandwidths by applying the Fourier transformation[34] is intrinsically more complex and subject to variation than the direct differentiation of the raw spectra, like that presented here.

The consistent upshift of the HAp bands evident from the derivative spectra, averaging at ~ 25 cm⁻¹ for the v₃(P-O) mode and ~ 20 cm⁻¹ for the v₄(O-P-O) mode compared to the same modes in ACP, presents the key spectral feature usable to distinguish HAp from ACP in an FTIR spectroscopic analysis. The collective upshift of all phosphate vibration modes upon crystallization of HAp from ACP can be explained by different effects, but the most probable and/or prominent in the case of a multifaceted mechanism is the larger amount of hydrogen bonding to phosphate groups in ACP than in HAp. As exemplified by water, the transition from water vapor to liquid water to hexagonal, atmospheric ice is accompanied by an increase in the average number of hydrogen bonds per water molecule, which results in the downshift of the symmetric stretch of the O-H bond from 3657 to 3385 to 3135 cm⁻¹, respectively[35,36]. Another effect at work here comes from the shortening of P-O bonds and a corresponding increase in their rigidity accompanying the transition to a more intense crystal field in crystalline HAp compared to ACP, the reason for which the phosphate bands are generally upshifted in HAp by ~ 20 cm⁻¹ compared to the free phosphate vibration modes. The same bond stiffening effects are responsible for the stretching vibration mode upshifts in numerous other materials and supramolecular structures, including citrates, which exhibit one such shift in their C=O stretching mode as they transition from a monomeric to a dimeric[37] or a trimeric[38] configuration. Based on the results of earlier thermogravimetric studies, the amount of water bound to ACP exceeds approximately three times the corresponding amount of water in Hap[39]. The more copious hydrogen bonding decreases the vibration frequency compared to that of the freer atomic groups accommodated in an ideal, water-free lattice. This critical effect of water agrees both with the fact that water release plays an intimate role in the transformation of ACP to Hap[40] and with a more general fact that dehydration at the surface is usually the most energetic step in the growth of a crystal from atomic or molecular precursors[41].

Fig. 5a shows the raw FTIR spectra in the 1200 – 500 cm⁻¹ range for ACP aged for different periods of time in its parent solution under ambient conditions and, per the diffractometric data shown in Fig. 2b, eventually transforming to HAp. Unlike the diffractometric data, which indicate a sudden transition between the two phases, a more gradual process is hinted at by the FTIR results. One indication for this comes from the similarity between the spectra of ACP after 5 and 60 min of aging and the similarity between the spectra of ACP after 120 and 180 min of aging, but only in the v₃(P-O) region of interest. In the v₄(O-P-O) region, ACP powder immediately removed from the solution exhibits more intense absorption than ACP samples aged for finite periods of time, but the peak features of all samples, aged or not aged, are indistinct, with no frequency shifts or shape differences being detected (Fig. 5a). After the first-order differentiation of the spectra is performed, the newly formed features in the v₃(P-O) region start to indicate a distinct difference of the unaged amorphous sample compared to the aged samples. The sub-regions in which this difference is most noticeable are marked with boxes 1–3 in Fig. 5b. After the second-order differentiation, these spectral differences in the v₃(P-O) region get somewhat diluted, but a definite shift becomes detectable in the v₄(O-P-O) region (Fig. 5c). Namely, both positive components of this band become upshifted as the aging time increases from 5 min to any larger duration, the less energetic one by 575.5 cm⁻¹ – 572.5 cm⁻¹ = 3.0 cm⁻¹ and the more energetic one by 616.3 cm⁻¹ – 613.8 cm⁻¹ = 2.5 cm⁻¹, presenting a key spectral feature for evidencing the early onset of the crystallization process in ACP aged in the solution using FTIR spectroscopy.

Fig.5.

FTIR spectra of an ACP powder isolated from the solution immediately after precipitation and the same powder aged under ambient conditions until transforming to HAp (a). First-order (b) and second-order (c) derivatives of the same FTIR spectra as in (a). Dashed rectangles in (b) denote regions where the spectral features of unaged ACP (5 min) significantly deviate from the spectral features of the aged precipitates (60 – 180 min).

The difference between the derivative IR spectra of ACP and HAp obtained by aging in the solution (Fig. 5) is clearly lesser than that detected in dry powders (Fig.4). One reason for this may be the effect of hydrogen bonding on shifts of the P-O vibration modes to lower frequencies in the amorphous phase. Namely, because both ACP and HAp in suspended or gelatinous states retain a large amount of bound and structural water, the effect of crystallinity on these shifts is lesser at the same order of differentiation compared to the effects in the dry state. Still, the differentiation of the spectra, regardless of the order, demonstrates the greatest degree of change in a material at the spectroscopically sensitive atomic level to occur at early aging time points and this is in contrast with the diffractometric data, which conform to the nucleation lag time model. Per this traditional model describing the crystallization of HAp from an amorphous precursor, this process is defined by the nucleation lag time, a preparatory period culminating in a rapid transition of the amorphous phase to the crystalline one[42]. According to the model, the preparation for this rapid transition can be lengthy and is inversely proportional to the supersaturation ratio in the solution[43]. It is also influenced by the ionic composition of the medium[44] and by the nature and the concentration of heterogeneous nucleation surfaces in contact with the solution[45]. The transition itself, however, is rapid and no gradual change in the ratio between the crystalline and the amorphous phase is normally experimentally observable. This mechanism is supported by the diffractometric data, which capture ACP or HAp, but not mixtures of the two phases, let alone gradual transitions across their different weight ratios. However, although diffractometric in nature, one such model is clearly not spectroscopic too because, as the spectroscopic results presented here show, at the short-range level of individual atomic groups, the transition does not exhibit a sudden character and the time points at which the most critical changes in the material occur per the results of the FTIR analysis (i.e., between 5 and 60 min of aging, Fig. 5c) do not coincide with the time points at which the most critical changes in the material occur per the results of the XRD analysis (i.e., between 120 and 180 min of aging, Fig. 2b). This discrepancy highlights an oft-forgotten premise, which is that the empirical models of physical phenomena are greatly defined by the measurement tools used to build them, given that different tools would lead to models with different features. It also refutes the incorrect premise often heard in the context of discussing nucleation and crystal growth of HAp, which is that the induction period is a “dead” period, during which the system simply waits for a crucial impetus to help it transcend the energy barrier and reach the crystalline state, when in reality, at the atomic level, this induction period may be as lively from the organizational standpoint as the one succeeding it.

Conclusion

This study demonstrated the augmentation of spectral line differences through FTIR spectra differentiation and the resulting ability to distinguish the amorphous phase in CPs from its crystalline products with a greater degree of precision than by comparing the raw spectra. Specifically, (a) the positive v₃(P-O) peak component appearing at 998.1 cm⁻¹ in the second-order derivative of the spectrum of TCP and coinciding with the troughs yielded by the amorphous phase was deduced as a key spectral feature applicable to distinguish crystalline TCP from ACP; (b) the upshift of the peaks originating from phosphate vibration modes in both first- and second-order derivatives of the spectra of HAp compared to the spectra of ACP, averaging at ~ 25 cm⁻¹ for the v₃(P-O) mode and ~ 20 cm⁻¹ for the v₄(O-P-O) mode, was deduced as the key spectral feature applicable to distinguish crystalline HAp from ACP; (c) the upshift of the doublet originating from the v₄(O-P-O) phosphate group vibration mode in the second-order derivative of the spectrum of ACP with its aging time in the solution was used to deduce the occurrence of the early crystallization process in the ACP precipitate. The study utilized a rigorous kinetic approach, focusing on the comparison of spectral features of amorphous precursors annealed in air at different temperatures and aged for different periods of time in an aqueous solution until transforming to one or a mixture of crystalline phases, including hydroxyapatite and α- and β-tricalcium phosphate. It demonstrated that the kinetics of the physicochemical changes deduced to be present in the amorphous samples undergoing crystallization is dependent on the measurement technique and is different in the spectrometric settings compared to the diffractometric ones. Thereby, the study challenged the concept of the nucleation lag time preceding the crystallization from amorphous precursors in CPs as a “dead” period and derived a finite degree of changes occurring during it at the atomic scale from the discrepancy between the spectroscopic and the diffractometric findings. Finally, the study is meaningful from the research standpoint, but also easily implementable in relevant instructional settings.