1 Introduction

For years, orthopaedic surgeons have been using many synthetic materials (metals, plastics, polymers, ceramics etc.) to replace human tissues in order to restore their lost functions. A biomaterial was defined as “any substance or combination of substances (other than a drug), synthetic or natural in origin, which may be used for any period of time, as a whole or as a part of a system which treats, augments, or replaces any tissue, organ, or function of the body” (A Consensus Development Conference on the Clinical Applications of Bio-materials. National Institutes of Health, Bethesda, Maryland, USA) [1]. The biomaterial market is expanding rapidly, parallel to the growing demands for bone graft substitute materials. The substitutes should promote the formation of new natural bone, while the external material should be degraded. Therefore, a great variety of specially synthesized calcium phosphate materials to be used as bone graft substitute materials is proposed. Although all these biomaterials are of the same family, there are differences between the product properties, e.g. crystallinity, porosity, phase purity, grain size, mechanical and thermal stability. The desired characteristic depends on the required application of the material.

Synthetic calcium orthophosphates may be used for hard tissue replacement due to the fact that they are analogous to the mineral part of calcified tissues [2, 3]. Hydroxyapatite (HA) Ca10(PO4)6(OH)2 is the major mineral phase in natural bone and teeth, so synthetic HA is widely used in medicine and dentistry because of its biocompatibility and bioactivity properties. This close relationship has been demonstrated by in vivo and in vitro studies [4, 5]. Beta–tricalcium phosphate (β-TCP) is one of the most popular bioresorbable materials for bone substitutes [6, 7]. Besides, a series of porous CaCO3/HA composites were also studied [8].

Producers of bone graft substitute materials often describe their products as chemically and mineralogically identical to the inorganic part of bone [9–11]. Because bone grafts, as well as their substitutes, are sterilized by ionizing radiation (electron beam or gamma source), one may expect that the radiation-induced changes after irradiation will be similar for natural and synthetic materials. The changes can be studied by electron paramagnetic resonance (EPR) spectroscopy.

The EPR phenomenon was discovered in 1945 by Zavoisky [12]. This unique non-destructive spectroscopic technique can detect all kind of paramagnetic species, which possess one or more unpaired electrons, namely, free radicals, biradicals, triplet excited states, defects in solid, crystal imperfection, and most transition metal and rare earth ions. These entities all play an important role in chemical and biological processes. An important category are radicals in solids, characterized by one unpaired electron. When a paramagnetic system is placed in a magnetic field, its energy levels split (Zeeman splitting). Resonance absorption of electromagnetic radiation energy is detected and recorded as first or second derivative of the absorption curve.

One of advantages of EPR spectroscopy is the possibility of determining the structure of radicals and EPR-active centers. This feature is mainly a result of the interaction between spin of unpaired electron and those of neighboring atomic nuclei with non-zero nuclear spin and it leads to hyperfine splitting of EPR lines. However, even if unpaired electron does not interact with magnetic atoms, it is possible to determine some properties of the vicinity using g-factor of EPR signal. g-Tensor is characteristic for a paramagnetic system and is determined by two factors. The first one is determined by the central atom or molecule containing the unpaired electron (free system) and the second one is affected by its nearest environment, mainly the host lattice and its defects and impurities. The principal values for paramagnetic system can be calculated either analytically by perturbation methods or numerically, using, for example, density functional theory (DFT) methods. A adequate formula of this type, applicable to many molecules has been derived by Stone [13]. Thus, the principal g-values contain some information about the electronic structure of the paramagnetic species.

EPR phenomena has been described in numerous books and papers, e.g. [14], [15].

The both synthetic and biological hydroxyapatites have been investigated for more than 50 years. At present this technique is used as a rapid, sensitive and accurate method for control of irradiated food, dating, accidental dosimetry and in medical studies to follow the remineralization of bone grafts and bone growth on scaffolds [15–19]. The EPR spectra of γ-irradiated hydroxyapatites are complex and consist of several signals due to the paramagnetic species, derived mainly from carbonate impurities in HA lattice. They may substitute the hydroxyl groups (the so-called A site), phosphate groups (B site) in crystalline lattice of hydroxyapatite or those located on the surface of microcrystallites. Paramagnetic species differing in the molecular structure and charge, the CO3 3−, CO3 −, CO2 −, O− and O3 −, have been identified by analysing their spectroscopic EPR parameters [20]. The most stable and intense EPR signal in both natural and synthetic apatites is an anisotropic singlet of an orthorhombic symmetry, with the following spectroscopic parameters of the g tensor: gx = 2.0030, gy = 1.9973, gz = 2.0017 and peak-to-peak width ΔHpp = 0.85 mT. Its intensity increases linearly with an increasing dose up to 20 kGy [21]. It represents the CO2 − radical anion located at the surface position [19, 20]. EPR spectroscopy allows usually to determine the nature of paramagnetic species, but the determination of the location of these species in the host lattice remains less evident because of the limited resolution of EPR spectra to determine of superhyperfine interactions with the neighbouring nuclei.

In synthetic hydroxyapatites, either precipitated from an aqueous solution and dried at 400 °C, or synthesized at high temperature, the most important carbonate derived radicals located in crystalline structure detected so far at room temperature (RT), are CO3 3− radicals substituting a phosphate or hydroxyl group [22–24]. The same radicals were identified directly after irradiation at 77 K and annealed to room temperature in human tooth enamel dried at 400 °C [25]. However, in bone and normal tooth enamel these signals could only be isolated so far after tedious computer manipulation of EPR spectrum.

The main goal of the present work is to study the paramagnetic species in irradiated bone substituting grafts in order to investigate selected materials with the biomimetic properties concerning radiation effects. Radiation sterilization of artificial materials which not only are to replace bone, but also undergo rebuilding to form new bone should be performed under the same conditions as for bone grafts. Thus, it is preferable that the radiation changes will be similar in bone grafts and substituting materials.

In the present study we apply X and Q band EPR spectroscopy to identify of the paramagnetic centers which are responsible for EPR signals, that are stable at room temperature in γ-irradiated bone substituting biomaterials based on hydroxyapatite. We have examined synthetic hydroxyapatities and two commercial synthetic bone graft substitutes. The results obtained have been compared with those for irradiated compact bone. The same set of samples was examined by FT-IR spectroscopy.

2 Experimental

Two types of synthetic bone substituting materials—Nanobone® (Artoss GmbH) and HA Biocer (CHEMA), a synthetic chemically pure hydroxyapatite powder Ca5HO13P3, purchased from Sigma-Aldrich Co. (No 289396) and powdered defatted human compact bone samples in natural form, obtained from the National Centre of Tissue and Cell Banking, Warsaw, Poland, were used in this study.

All samples were irradiated at room temperature at the Institute on Nuclear Chemistry and Technology, Warsaw, Poland, with a dose of 5 kGy in a 60Co gamma source “Issledovatel”, with a dose rate of 0.8 kGy/h. Then, the samples were powdered and placed in thin-wall spectrosil quartz tubes.

The EPR measurements at X- (9.5 GHz) and Q-band (34 GHz) were carried out at room temperature just after the gamma irradiation, 1 day and 5 days later, in order to make sure that all signals derived from the unstable radicals decayed. The X-band spectra were recorded with the Bruker ESP-300 spectrometer, whereas the Q-band measurements were carried out with the Bruker ELEXSYS E-500 spectrometer. Our X and Q-band spectrometers are equipped with precise frequency counters and gaussmeters. For both bands, a wide range of microwave powers and modulation amplitudes were tested in order to optimize the detection conditions.

The measurement parameters were the following: for the X-band, the modulation frequency was 100 kHz, the modulation amplitude was 0.1 mT, and the microwave power was 1–10 mW. For the Q-band modulation, the amplitude ranged 0.05–0.1 mT, while the microwave power ranged 0.1–2 mW. A standard polycrystalline DPPH sample (g = 2.0036) was used for an accurate g-tensor determination.

Spectroscopic FT-IR studies were carried out at 298 K on a Perkin Elmer Spectrum 1,000 spectrometer. The absorption spectra were acquired with 2 cm−1 spectral resolution from KBr pellets using a MCT detector and 50 scans. They were processed using the GRAMS/AI 8.0 software (Thermo Scientific, 2006).

3 Results

3.1 EPR measurements

3.1.1 Compact bone

The EPR spectrum of compact bone powder recorded at the X band 1 day after irradiation is presented in (Fig. 1a). This spectrum is a complex one with a dominant anisotropic singlet of an axial anisotropy, with the spectroscopic g factors of g⊥ = 2.003 and gII = 1.997, representing the CO2 − radical anions located on the surface of the hydroxyapatite crystallities. Other signals, with much lower intensity, overlapping the CO2 − singlet, most probably represent the carbonate radical anions CO3 − stabilised at different sites, with giso = 2.009 and 2.012. After a few weeks of storage, the carbonate radical anions disappeared, and only the very stable signal of CO2 − radical was detected. This signal, very characteristic for bone samples, did not show any noticeable changes of the shape and intensity for years [26, 27].

Fig. 1
figure 1

EPR spectra of compact bone powder sample γ irradiated at room temperature a recorded in X band 1 day after irradiation b recorded in Q band 1 day after irradiation

Full size image

The spectra recorded in the Q band are better resolved, which makes them easier to identify the individual paramagnetic species in complex EPR spectrum. The Q band spectrum of the irradiated compact bone (Fig. 1b) reveals that the CO2 − radical anion has orthorhombic symmetry, with characteristic spectroscopic g factors: gx = 2.003, gz = 2.0015 and gy = 1.997—with agreement with earlier results [26].

3.1.2 Synthetic hydroxyapatite

The EPR spectrum of synthetic, chemically pure hydroxyapatite powder purchased at Sigma-Aldrich Co, recorded at the X band, shows the anisotropic singlet characteristic for the CO2 − radical anion (Fig. 2a). The signals with much lower intensity, visible in the low-field part of the spectrum, derive from CO3 3− and CO3 − radicals and/or impurities in the sample. The EPR spectrum of the same sample recorded at the Q band of much better resolution proves that it consists of two separate signals (Fig. 2b). The first one is similar to the signal recorded in human bone and represents the CO2 − radical anions, with the following spectroscopic g parameters: gx = 2.003, gz = 2.0014 and gy = 1.997. The second one is an isotropic singlet with giso = 2.0006. An identical signal was recorded in natural and synthetic aragonite matrices and was assigned to the CO2 − radical formed in the domains containing occluded water [28, 29].

Fig. 2
figure 2

EPR spectra of synthetic hydroxyapatite powder sample γ irradiated at room temperature a recorded in X band 1 day after irradiation b recorded in Q band 1 day after irradiation

Full size image

3.1.3 NanoBone®

NanoBone® is a synthetic bone substitute or bone reconstructing materials used for dental and bone surgery. It consists of 76 % nanocrystalline hydroxyapatite embedded in a silica matrix (24 % SiO2 in weight). The material is formed as granules in the dimensions of 0.6–1 mm × 2 mm. The nanocrystalline HA is similar to the hydroxyapatite of the bone. The silica promotes the formation of collagen and bone tissue, when the graft is surgically placed into living organism. Nanobone® is biocompatible and completely substituted by bone during the process of bone rebuilding [30].

The EPR signal of the irradiated synthetic bone graft substitute Nanobone® detected in the X band (Fig. 3a) looks like a broad singlet with two inflection points. The Q-band spectrum (Fig. 3c) shows clearly the signal anisotropy with the orthorombic g-factor of gx = 2.0041, gy = 2.0036, gz = 2.0018, characteristic for the CO3 3− anion radical formed as the result of the electron capture by the CO3 2− anion situated at the phosphate site in crystalline lattice [20]. Additionally the spectra at the X- and Q-band show very weak feature at a high field, with the g-factor of 1.9973, representing the gll component of the CO2 − signal [20]. The perpendicular component of the CO2 − is barely visible because it overlaps with the CO3 3− signal. After 5 days, the EPR spectra reveal the significant changes (Fig. 3b). The signal assigned to the CO2 − becomes more clearly visible, showing an inflection point at g value equal 2.0026 characteristic for g⊥ of CO2 − radical anion. Similar changes in the intensities of the CO3 3− and CO2 − radicals were recorded in tooth enamel heated to dry mass and X-ray irradiated at 77 K. After thermal annealing at 0 °C, the EPR spectrum consisted almost entirely of the CO3 3− signal but as a results of storage at room temperature for 2 days the relative contribution of the CO2 − signal increased due to the partial decay of the CO3 3− radicals [25].

Fig. 3
figure 3

EPR spectra of Nanobone® powder sample γ irradiated at room temperature a recorded in X band 1 day after irradiation b recorded in X band after 5 days storage c recorded in Q band 1 day after irradiation

Full size image

3.1.4 HA Biocer

HA Biocer is a pure hydroxyapatite powder similar to the bone mineral, so-called bone apatite. It is compatible with teeth and bones of the human body [10].

The EPR signal of the irradiated HA Biocer detected at the X band is shown in Fig. 4a. The main component of the EPR spectrum is an anisotropic signal with an orthorombic symmetry and the g-factor of gx = 2.0042, gy = 2.0036, gz = 2.0018. The g-values are similar to the signal of the NanoBone® sample, which was assigned to the CO3 3− at the phosphate site.

Fig. 4
figure 4

EPR spectra of HA Biocer powder sample γ irradiated at room temperature a recorded in X band 1 day after irradiation b recorded in X band after 5 days storage

Full size image

After 5 days of storage, the EPR spectrum is changed in a similar way as in the NanoBone® sample, showing a weak high-field component (gll = 1.997) of the CO2 − signal (Fig. 4b).

The Q-band EPR spectrum recorded 1 day after the irradiation very much resembles the spectrum of the NanoBone® material at the same frequency.

3.2 FT-IR measurements

Biocer, NanoBone® and HA Aldrich gave IR spectra typical of apatites, with dominating phosphate bands in the ν1ν3-PO4 3− (900–1200 cm−1) and ν4-PO4 3− (500–700 cm−1) regions (Fig. 5). In the NanoBone® spectrum there were clearly discernible extra bands of silica at 812 and 469 cm−1 from symmetric stretching and bending vibrations, respectively, of the Si–O–Si bonds. Carbonate bands were hardly visible and only B-carbonates have been detected. Using an intensity ratio of the 1417 and 603 cm−1 bands [31, 32], we have estimated the B-carbonate content in those materials at ca. 0.5 wt%. For HA Aldrich, a distinct band was found at 868 cm−1, too big to be from a ν2-CO3 2− vibration. We have assigned it to HPO4 2− ions. Crystallinity of those materials was assessed by observing resolution of the phosphate bands (1090 vs. 1033 cm−1; the 603 cm−1 band against its neighbours; sharpness of the 962 cm−1 band) and by looking at the intensity of the 3572 cm−1 band from structural OH− uninvolved in H-bonding. Accordingly, the apatite crystallinity decreased in the series: Biocer > NanoBone® > HA Aldrich. The water content increased in the order: Biocer < NanoBone® < HA Aldrich, as found from the stretching and bending water bands at 3430 and 1640 cm−1, respectively.

Fig. 5
figure 5

Comparison of the FT-IR spectra of Biocer, NanoBone® and HA Aldrich. The absorbance has been standardized to the highest band at 1033 cm−1

Full size image

The defatted bone gave much different IR spectrum (Fig. 6), which comprised both apatite and collagen bands. The water bands were huge comparing to the formerly discussed samples. Unfortunately, weak carbonate ν3-CO3 2− bands have been obscured by stronger collagen bands. However, the 870 cm−1 band of the v2-CO3 2− vibration appeared much higher between the phosphate bands in comparison to the other materials, indicating the highest carbonate content in the defatted bone. Certainly, this sample contained B-carbonates, as can be confirmed by the 1417 cm−1 band. Nothing can be said about A-carbonates. Then, the apatite of this sample had the lowest crystallinity of the studied materials, as evidenced by the worst resolution of the ν1ν3-PO4 3− domain. The bands of the latter region were so broad that it appeared with particularly low peak intensity comparing to the ν4-PO4 3− region. As well, the ν4-PO4 3− domain in the defatted bone IR spectrum was the least resolved of all the studied cases.

Fig. 6
figure 6

Comparison of the FT-IR spectra of defatted bone and pure collagen. The absorbance scale in both spectra has been adjusted in order to conform the 2924 cm−1 bands (C–H asymmetric stretching vibration in –CH2−)

Full size image

4 Discussion

Both materials commercially available as bone substitutes, which were investigated in our study, have been already applied in clinics. Good clinical results of using HA-Biocer in periodontal diseases as well as in the treatment of enamel hypoplasia were reported [33–35]. In the study reported by Zietek et al. it was compared to the other bone fillers used in periodontal surgery [35]. The results were assessed after 8 years basing on clinical and radiological observation. Satisfactory results were found for all materials used, i.e. Bio-Oss, Biogran and HA-Biocer. With NanoBone®, promising clinical results of using this material for maxillary sinus floor augmentation have been only recently reported [36]. Histologically confirmed NanoBone® conductivity after jaw bone sites augmentation was shown as well [37]. On the contrary, only limited horizontal ridge width alterations following tooth extraction was found in a randomized clinical trial, where NanoBone® was applied in comparison to demineralized bone matrix [38]. The results for both materials were similar. Generally, it is not easy to form an opinion about various materials used for tissue substitution basing on clinical results. Data come from various types of surgery and observation periods. They involve various sets of materials and are based on different methodology. For obvious reasons a wide comparison of various materials in one surgical system is practically unattainable. What is worse, even experimental studies performed under well-defined conditions not always bring univocal results. In case of NanoBone®, unsatisfactory results of osteoblast-like cells response toward the material in respect of cell adhesion and proliferation in culture was reported by Bernhardt et al. [39]. At the same time, good cytocompatibility of NanoBone®, in reference to BioOss®, which is commonly accepted as one of the best tolerable bone fillers, was shown in culture of human bone derived cells by Liu et al. [40]. By intuition, the more bone-mimetic material, the better potential bone substitute might be expected. The majority of synthetic bone grafts, both candidates and those which have been already in clinical usage are chemically similar to bone mineral. Two commercially available bone substitutes investigated in our study are hydroxyapatites. However we have shown that they missed the bone specific EPR characteristics after irradiation. On the other hand, we have found that the EPR spectrum of the third investigated material, which was hydroxyapatite of no medical grade, purchased in the chemical company, is most similar to the spectrum of bone hydroxyapatite. It indicates that it is possible to obtain synthetic materials which mimic bone mineral in respect to the EPR characteristics after irradiation. As for now, we do not know if this phenomenon may be connected with material capacity toward bone regeneration. However this is a new finding which seems to be worth of further investigations.

From spectroscopic point of view, the appearance of the CO2 − EPR signal during the storage of both bone graft substitutes is the most striking result of our studies. It seems reasonable to consider two mechanisms. In the first one, there is a transfer of an electron from the CO3 3− located inside the HA microcrystals towards the CO2 molecule adsorbed on the surface of crystallites, leading to the formation of CO2 −. In order to prove this hypothesis, quantitative EPR studies are necessary for much longer sample storage. Then the deconvolution of broad CO3 3− singlet overlapping the perpendicular component of the CO2 − anisotropic singlet might prove that the increase of the CO2 − intensity is related to the decay of CO3 3− radical.

The other hypothetical mechanism might be related to the photoionization of the CO2 surface molecules by sunshine UV. Although this process seems much less probable than the first one, we are going to verify it by comparing the spectral changes in samples kept in dark and in samples exposed to UV irradiation from a mercury lamp.

The EPR results prove clearly that the long-lasting radiation effects associated with carbon-centered radicals in commercial bone substitutes are different from those in biological hydroxyapatites. It might be related to the different locations of the CO3 2− anions and/or the CO2 molecules in both types of materials.

5 Conclusions

HA Biocer and NanoBone® are commercial synthetic bone substituting or reconstructing materials based on hydroxyapatite. Producers of both graft materials assure that the hydroxyapatite used in their product is almost the same as in autologous bone. However, the EPR study of irradiated compact bone and the synthetic graft materials suggests that their microscopic structures are different. Although radiation induces the same types of anion radicals in both types of materials, but their populations differ very much. In irradiated bone, the dominant paramagnetic centers represent the CO2 − radical, whereas the other radicals, as CO3 − and CO3 3−, have much lower concentration. In HA Biocer and NanoBone®, after irradiation the CO3 3− radical anions are the major paramagnetic products with a much higher concentration than the CO2 − radical anions.

Unexpectedly, the population of the CO2 − radicals grows during sample storage at room temperature. To explain this effect additional experiments will be undertaken.

Among the studied commercial bone substituting materials the radiation effects in Sigma-Aldrich hydroxyapatite most resemble the effects in bone hydroxyapatite. In both minerals the CO2 − radical anion is the major paramagnetic product of radiolysis at room temperature.

The EPR study of irradiated compact bone and the synthetic graft materials suggest that their microscopic structures are different.

In FT-IR spectra of NanoBone®, HA Biocer and synthetic hydroxyapatite the PO4 3− bands are the most intense, whereas in compact bone bands due to collagen dominate. CO3 2− groups located in B-site are detected in synthetic materials. CO3 2− groups located in A-site are not visible in FT-IR spectra. It is in agreement with EPR results, where signals derived from CO3 3− radicals located at A-site are not detected, while signal due to CO3 3− radicals located at B-site dominates in HA Biocer and NanoBone® spectra.

The results presented prove that EPR spectroscopy is a useful tool for studying subtle structural changes in bone substitute materials.