1 Introduction

Silicate-based and borate-based bioglass of nominal compositions (45SiO2, 24.5CaO, 24.5Na2O, 6P2O5 wt%), (45B2O3, 24.5CaO, 24.5Na2O, 6P2O5 wt%), respectively, containing variable Cu contents (x) in the range (0.2–4 additive) have attracted substantial interest due to their remarkable potential for use in health applications, e.g., in bone bonding [1, 2]. Thanks to the antimicrobial properties of copper, bioactive glasses containing Cu are useful in medical applications where preventing infections is important [1,2,3]. When the glass comes in contact with bodily fluids, Cu ions are released, which can help inhibit the growth of bacteria and other microorganisms and increase the ability to promote tissue regeneration and its compatibility with biological systems [1,2,3,4,5]. Also, glass samples doped with Cu ions are used in wound healing applications as they have been shown to drastically boost blood vessel formation [6]. Borate glass is known for its unique behavior of fast degradation and formation of hydroxy carbonate apatite (HCA) and excluded with urine in addition to low cost and easy processability [1]. The reason for such a fast transformation was completely missed [1, 2]. The formation of hydroxyapatite (HA) or hydroxycarbonate apatite (HCA) on the surface of bioactive glasses and glass–ceramics is a crucial phenomenon that determines their bioactivity and bone-bonding ability [1,2,3,4]. Cu ions, in both their oxidation states, Cu (I) and Cu (II), can play a significant role in promoting HCA formation during the preparation of borate bioglass doped with Cu [7]. Further, Cu (II) ions can act as nucleation sites for the formation of HCA due to their ability to interact with the functional groups present in the surrounding environment since the presence of Cu (II) ions can enhance the formation of amorphous calcium phosphate (ACP) precursors, which serve as intermediates in the formation of HCA. Moreover, Cu (II) ions can also participate in the dissolution of the bioglass matrix, releasing calcium and phosphate ions, which are essential building blocks for HCA formation. Monovalent Cu (I) ions can also contribute to the formation of HCA, although their role may be less pronounced compared to Cu (II) ions. Cu (I) ions can participate in redox reactions, leading to the formation of reactive oxygen species (ROS), which can influence the local environment and promote the nucleation and growth of HCA. Cu (I) ions may also interact with functional groups on the bioglass surface, facilitating the adsorption and deposition of calcium and phosphate ions, leading to HCA formation [7,8,9].

To understand the role of Cu incorporation modifier silicate and borate glasses, X-ray absorption fine structure spectroscopy (XAFS) based on synchrotron radiation was used to explore the oxidation states as well as the local structure and coordination of Cu atoms [2,3,4,5,6]. Generally, to explain the specific physical and chemical properties of glassy materials, it is necessary to probe deeply into the structural properties of these materials; however, glassy materials are very often either amorphous or semi-crystalline materials characterized by the absence of long-range order [10]. The structural characterization of glassy materials has always been performed by probing the local atomic structural environment in the short-range order (1–5 Å) [11]. The challenge is then to investigate the effect and possibly establish a correlation between the short-range structure and the resulting different properties and applications of the bioglass materials. The high relevance of the structural investigations of glassy materials is not only due to the interest in explaining the physical and chemical features of these materials but also due to their wide spectrum of daily and potential applications [12]. Glass can be used, in medical applications as in bone bonding as mentioned above, in energy applications as in fuel cells, and in storing radioactive wastes [1, 2].

For more details, being a short-range local structural and selective elemental technique, XAFS is an ideal tool to structurally describe glass materials in the absence of long-range ordered structure [13]. XAFS provides valuable local structural information about oxidation states [14, 15], local coordination geometry, site-symmetry [16, 17], kind/type of neighbouring atoms, relative disorder, and angles between the central atom and nearest neighbours [18, 19]. Unlike some other techniques, XAFS can be used for probing either bulk or surface structure, by tuning the XAFS data collection mode [20]. Also, XAFS does not rely on bulk properties but is sensitive to atomic structure. XAFS operates at core-level X-ray absorption energies, corresponding to specific electron shells (typically, K and L shells). Also, due to its high sensitivity to subtle changes in the local environment, XAFS can detect variations in bond lengths, coordination numbers, and oxidation states corresponding to the change in composition/concentration of the central atom.

Generally, the XAFS spectrum consists of two energy regions: near-edge XAFS (also called X-ray absorption near-edge spectroscopy, XANES) that denotes the spectral region around the absorption edge, and extended XAFS (EXAFS) starting from around 100 eV after the absorption edge and extended up to more than 1000 eV. XANES probes the density of either the empty or partially filled electronic states by exciting the inner shell electron to those states [21]. As this electronic transition may only occur if it is allowed according to the dipole selection rule [(Δl) is ± 1], XANES probes the angular momentum of unoccupied electronic states.

On the other side, extended XAFS (EXAFS) is sensitive to the surrounding local structure of the absorber atom. The origin of the EXAFS signal is the interactions between the photoelectron excited from a deep core level of the absorbing atom by an incident X-ray photon with the electrostatic potentials of neighboring atoms. The presence of neighbouring atoms slightly perturbs the absorption probability, resulting in the oscillatory structure of the absorption coefficient μ(E) on top of smooth variations of the absorption probability μ0(E) that is unrelated to the sample structure. EXAFS delves deeper into the atomic structure, revealing details about neighbouring atoms and their bond distances [22, 23].

The aim of this work is to determine the oxidation states and local structural details of Cu atoms in order to probe the effect of boosting Cu oxide at the expense of B2O3 and SiO4 in borate and silicate bioglass samples, respectively. The atomic local structure was studied by collecting Cu K-edge (8979 eV) XAFS spectra at beamline BM08-XAFS/XRF at the SESAME synchrotron [24, 25]. As stated before, for the healing process in bone bonding applications, the coexistence of Cu2O and CuO within silicate or borate samples, with a significant wt% for each phase, will promote the conversion process to allow for a faster formation of HCA. Consequently, this work intends to find out the optimum Cu content (x), which forms a mixture of Cu (I) and Cu (II) with a significant wt% for both Cu phases.

2 Experimental procedure

2.1 Sample preparation

Silicate-based and borate-based bioglass of nominal compositions (45SiO2, 24.5CaO, 24.5Na2O, 6P2O5 wt%), (45B2O3, 24.5CaO, 24.5Na2O, 6P2O5 wt%), respectively, containing variable Cu contents (x) in the range (0.2–4 additive) were synthesized using the melt quenching technique. Sample names and compositions are listed in Table 1. Analytical grade copper oxide was used as received, while calcium and sodium oxides were obtained from their carbonates. Ammonium dihydrogen phosphate and orthoboric acid were used as a source of phosphorus pentoxide and boron oxide, respectively. Weighed batches were sintered at 450 °C to remove carbonate, ammonia, and water. The oven temperature was gradually raised to 1200–1300 °C depending on the glass composition. The obtained melts were swirled many times to obtain a bubble-free sample. The melts were then poured into stainless steel moulds of the required dimensions.

Table 1 Silicate and modified borate bioglass compositions (in wt%) at different Cu content (x)
Full size table

To prepare the silicate and modified borate samples containing Cu for XAFS measurements, the calculated weight from each sample powder was homogeneously distributed into a reasonable amount of polyvinylpyrrolidone (PVP). The sample/PVP composites were pressed into pellets of 13 mm in diameter to allow the beam to probe a large area of the sample.

2.2 XAFS measurements and data collection

Using the BM08-XAFS/XRF beamline of the SESAME Synchrotron facility in Allan-Jordan [24, 25], Cu K-edge XAFS spectra were adequately acquired at ambient room temperature. SESAME is operated at 2.5 GeV with a maximum injection current of 250 mA [25]. The BM08-XAFS/XRF beamline has been constructed for providing hard X-rays in the range of (4.7–30) keV with high photon flux (109–1012 ph/s) and high beam stability and is dedicated to XAFS and XRF spectroscopic applications [24]. The beam spot size at the sample position is (1–20) mm horizontally, while it is (1–5) mm vertically. Double crystal monochromator equipped with Si (111) crystals is utilized, achieving an energy resolution ΔE of ~ 2.4 × 10–4 eV. Higher harmonics are effectively suppressed thanks to the Si-coated stripe of the collimating and focusing mirrors. Cu (0) metal foil was used for beamline energy calibration by assigning, through the derivative, the first inflection point of the tabulated K-edge of Cu foil (8979 eV). The sample XAFS spectra were collected in fluorescence mode (FL-mode) by using a silicon drift detector (SDD: Ketek) at 45 degrees with respect to the sample. The incident intensity (Io) was measured using a 150 mm ionization chamber (IC). The mixture of He, N2, and Ar gas in the IC was optimized for a ~ 15% absorption of the total incident beam. XAFS data for reference samples (CuO, and Cu2O) were collected in transmission mode where the ionizing gas (He, N2, and Ar) in ionization chambers for sample transmission (I1) and reference (I2) was optimized for an absorption of 65% and 90%, respectively.

Our XAFS data were collected in the FL-mode with an energy step of 0.5 eV in the pre-edge region, 0.2 eV in the XANES region, and with a constant k-step of 0.03 Å−1 in the EXAFS region. Four scans for each sample were recorded to improve the signal-to-noise ratio.

2.3 XAFS data processing and analysis

XAFS data were processed using standard procedures offered by the ATHENA program of the DEMETER software package [26, 27]. Accordingly, pre-edge absorption removal, background subtraction, normalization, energy conversion into k, weighing scheme, and Fourier filtering were performed. The threshold energy (E0) was taken at the first inflection point of the absorption edge jump, also, initially, it was taken at the fraction of the edge step (at 0.5 of the normalized intensity). Also, the white-line maximum peak energy was determined. Linear combination fitting approach (LCF) was performed for each sample considering Cu2O and CuO as constituent phases. Different XANES spectral features parameters were quantified from the XANES fingerprint spectra of different samples along with those of Cu standard compounds.

Fourier transforms (FT) of the k2-weighted EXAFS function were taken by using a Hanning window. Artemis software package (Artemis program2 from DEMETER; version 0.9.26), was used for modeling and structural shell-fitting of the collected EXAFS oscillations. The analysis of the experimental EXAFS has been done by generating the theoretical model and fitting it to the collected data. The best fits of the EXAFS signal were done in R-space in the interval of 1.35–2.117 Å with Hanning window in the range of (3–9.6) Å−1 of k. Only the first coordination shell was subject to the fitting since there is no clear spectral features for the next higher coordination shells due to/confirming the highly amorphous/disordered structure of the target glass samples.

Based on the results extracted from XANES fingerprint and LCF approach, the best fit structural parameters were extracted from EXAFS fitting by applying one structural model of Cu2O in case of silicate samples and two structural models of both Cu2O and CuO to fit modified borate samples. Considering the well-known EXAFS equation described in [23], the fit parameters included the amplitude reduction factor (S02) and the correction to the difference in photoelectron energy (E0) between the experimental data and theoretical model calculated using FEFF software package [27], which were set similarly for all scattering paths and fixed for all Cu content (x) to minimize the number of variable parameters and thus reduce the correlation among the fit parameters, also for a real comparison of the same samples with different x. While, interatomic distances (R) and variation in the interatomic distances or mean-square relative displacements (MSRDs), also known as Debye–Waller factors (σ2), were refined to get the best-fit result. S02 was determined to be ~ 0.97 by fitting the EXAFS data of the reference Cu foil with its well-known structure.

3 Results and discussion

3.1 XAFS study of silicate-based bioglass doped with Cu

Figure 1 exhibits the entire collected and normalized Cu K-edge XAFS spectra of bioglass of nominal composition (45SiO2, 24.5CaO, 24.5Na2O, 6P2O5 wt%) with variable Cu content (x) in the range of (0.2–4 additive). Figure 1 consists of two different XAFS parts; the first part is the near edge-XAFS or XANES, and the second part is the extended-XAFS or EXAFS. The two XAFS spectral parts are analyzed separately to extract complementary information about the electronic and local structure around Cu atoms. Consequently, to investigate the energy peak positions and spectral features of the pre-edge peak as well as the white-line of the XANES part, fingerprint XANES spectra of the samples are plotted along with those of Cu standard reference compounds as shown in Fig. 2.

Fig. 1
figure 1

Normalized Cu K-edge full XAFS spectra of Hench bioglass of nominal composition (45 SiO2, 24.5 CaO, 24.5 Na2O, 6P2O5 wt%) containing variable concentrations of copper oxide in the range (0.2–4 additive). All spectra are plotted with vertical offset, for clarity; xμ (E) is the total absorption, where x here is the thickness of the sample

Full size image
Fig. 2
figure 2

Normalized fingerprint XANES spectra of the silicate glass samples along with those of different Cu standards without offset (a), and with vertical offset (b)

Full size image

The extracted spectral white-line peak position, absorption energies taken as the first inflection point of the rising absorption edge, and nominal oxidation states of Cu cations in the silicate glass matrix are listed in Table 2. From the first look at the pre-edge feature and based on the extracted data in Table 2, it is clear that the XANES spectra of silicate samples are quite similar to that of the Cu2O standard reference. To confirm if the Cu cations in silicate glass form only a single phase of Cu2O or if there are other (secondary) phases/species of Cu compounds such as Cu-metal and/or CuO present, a linear combination fitting (LCF) was performed for each sample XANES spectrum. This way, the contributing ratio of Cu (I), Cu (0), and/or Cu (II) phases in different silicate glass samples can be determined. The results of the best fits to the data are tabulated in Table 2. It is observed that all silicate glass samples contain mainly Cu2O as shown in Fig. 3, which represents the best fit of silicate glass at x = 4 with Cu2O standard reference. This result matches well with the published results by Gaur et al. [28].

Table 2 Spectral white-line position, absorption energy value, taken at the first inflection point of the absorption edge jump, and oxidation states
Full size table
Fig. 3
figure 3

Exemplary LCF of silicate glass at x = 4 with those of Cu2O and CuO standard references

Full size image

EXAFS analysis was utilized to further explore the local structure around the Cu atoms. The Fourier transform (FT) is known as a complex function of bond length or distance R. The amplitude of FT is denoted by the real function χ(R); | χ(R)|. The peak positions in χ(R) are related to bond lengths between the Cu and neighbouring atoms, whereas the height of each peak is proportional to the coordination number N [10, 11, and 14]. However, we cannot directly extract the bond lengths and coordination numbers from χ(R). Figure 4 displays Fourier Transform magnitudes of the extracted k2-weighted EXAFS signals in silicate glass samples along with those of the Cu standard compounds (Cu (0), Cu (I) and Cu (II)) with y-offset (left) and without any offset (right), respectively.

Fig. 4
figure 4

k2-weighted EXAFS oscillations of silicate glass samples along with those of Cu standard compounds; shifted vertically for clarity (a), and without vertical offset (b)

Full size image

It is clear that the silicate sample doped with Cu exhibits a decrease in FT amplitude upon increasing Cu content (x). Accordingly, the coordination number N decreases, which means that the structure disorder increases [15].

The fit of the EXAFS data of Cu doped silicate samples is exemplarily displayed in Fig. 5 for the Cu content of x = 4. The extracted fit parameters for all silicate glass samples are tabulated in Table 3. The first coordination shell is formed by bridging the Cu cation between two oxygen atoms, where two oxygen anions are located at a mean distance of an average of 1.85 ± 0.02 Å from the absorber atom (Cu). Closer examination reveals that the second peak in the FT plot is significantly broader than the corresponding peak of the Cu (I) standard, which indicates a lower degree of crystallinity of the copper oxide structure formed in silicate glass samples [29, 30]. Although glass materials are generally amorphous according to X-ray diffraction (XRD) analysis, XAFS (EXAFS) probes the local short-range structural order and may, as such, identify local bonding geometries as found in CuO or Cu2O crystalline materials within an otherwise amorphous matrix.

Fig. 5
figure 5

Exemplary EXAFS fit obtained for the FT magnitude (upper) and the imaginary part (lower) of EXAFS signal at the Cu K-edge collected at x = 4 of silicate glass. The y-axis shows only the FT magnitude

Full size image
Table 3 The best-fit local structural parameters of the first coordination shell around Cu ions (Cu–O) within silicate glass samples; k = 2–9.6 (Å−1), R = 1–2.11 (Å), Hanning window, R-factor = 0.06 ± 0.02
Full size table

It was reported that the incorporation of Cu (I) in glass materials by ion exchange or by introducing to the batch, a two-fold coordination state was proposed, while other researchers [31,32,33] assumed a six-fold coordination number. In this work, the guessed coordination number of Cu–O in the first coordination shell of silicate glass samples was set at 2. This value fitted well the EXAFS silicate data with that of Cu2O standard and matched well with the fit performed by Tröger et al. [34].

3.2 XAFS study of the modified Borate-based bioglass doped with Cu

Figure 6 displays the normalized Cu K-edge full XAFS spectra of borate-based bioglass of nominal composition (45 B2O3, 24.5 CaO, 24.5 Na2O, 6P2O5 wt%). The extracted spectral white-line peak position, absorption energies, and nominal oxidation states of Cu cations in the borate glass matrix are shown in Table 2. It is observed from Table 2 that the white-line peak position (8996 eV) and the absorption edge (8985 eV) are diagnostics of mainly Cu (II) compounds, as reported in [31].

Fig. 6
figure 6

Normalized Cu K-edge full XAFS spectra of Hench bioglass of nominal composition (45 B2O3, 24.5 CaO, 24.5 Na2O, 6P2O5 wt%) with vertical offset

Full size image

Figure 7 shows XANES fingerprints of different borate glass samples along with those of Cu (0), Cu (I), and Cu (II) standard XANES spectra. XANES analysis shows that there is a small shoulder around 8879.5 eV, indicating the existence of the Cu (II) phase. On the other hand, all XANES spectral features and parameters in Table 2 show that the Cu cations in borate glass samples may take multiple oxidation states, i.e., Cu (II), Cu (I), and/or Cu (0) [34].

Fig. 7
figure 7

Normalized Cu K-edge XANES spectra of Hench bioglass of nominal composition (45 B2O3, 24.5 CaO, 24.5 Na2O, 6P2O5 wt%), along with those of Cu2O and CuO reference standard compounds, without offset (a) and with vertical offset (b)

Full size image

Figure 8 represents LCF model example at x = 4 of borate glass samples. It is observed that the different Cu cations with different oxidation states and their contribution within borate glass samples can be distinguished. The best fit to the data from the LCF approach are presented in Table 2.

Fig. 8
figure 8

Best fit results of LCF for Cu K-edge XANES spectrum of borate bioglass at x = 4 (as an example), along with those of Cu2O and CuO standards; LCF fitting in the range from − 20eV to + 20 eV around the absorption edge

Full size image

It is clearly shown that XANES spectral features and parameters of the modified borate glass samples are quite different from those obtained from the silicate glass samples. In order to confirm the difference between XANES spectral features of silicate and borate samples, XANES spectra of borate and silicate glasses at x = 2, (as an example), compared with those of XANES spectra of Cu standard compounds are shown in Fig. 9. For comparison, the borate glass XANES spectrum shows a shoulder at the absorption edge and an intense white line. While, the silicate glass samples show a characteristic sharp absorption peak without any pre-absorption peak like Cu (I) species due to its d10 electronic configuration, and thus lack of a classic 1s → 3d pre-edge transition.

Fig. 9
figure 9

Comparison between XANES spectra of borate and silicate bioglass at 2% of Cu (as an example), along with those of Cu standard compounds

Full size image

Figure 10 shows the FT magnitudes of the k2-weighted EXAFS signals of different borate glass samples (45 B2O3, 24.5 CaO, 24.5 Na2O, 6P2O5 wt%). The FT features of the modified borates can be modelled by considering the two structural models of the Cu (I) and Cu (II) standard references with different wt% (in the first coordination shell), as shown in Fig. 11. However, at longer distances, it seems that the local structure of borate glass samples has a higher disorder beyond the first coordination shell. It is clear that there is a remarkable reduction in FT amplitude with corresponding enhancement in the amount of Cu species. However, the samples of x = 0.2 and x = 1 present similarities in FT amplitude as well as the interatomic distance values. This result matches well with the LCF results Table 2. The reduction in the FT amplitude with changing Cu content is related to the evolution of Cu2O/CuO ratio, which leads to boosting the disorder parameter, which, in turn, leads to a lowering in coordination number N. Furthermore, the highest next coordination shells of Cu–O and Cu–Cu bonds are not as well seen as those for Cu2O and CuO standards, providing a lower degree of crystallinity of Cu-doped modified borate glass systems [35].

Fig. 10
figure 10

k2-weighted EXAFS signals of modified borate bioglass samples along with those of Cu standard compounds; shifted and arranged vertically (a) and without y-offset (b)

Full size image
Fig. 11
figure 11

EXAFS fit obtained for FT of EXAFS signal at Cu K-edge collected at x = 4 (as an example), of modified borate bioglass

Full size image

Table 4 clearly illustrates the fitted R-distances of the samples of x = 0.2, 1, and 2 are very close to each other, containing 2 (Cu–O) distances at 1.89 ± 0.02 Å, plus 6 (Cu–O) distances at 1.94 ± 0.02 Å. However, the sample of x = 4 shows a small change in the R-distances; 2 (Cu–O) distances at 1.91 ± 0.02 Å, plus 6 (Cu–O) distances at 1.95 ± 0.02 Å.

Table 4 The best-fit local structural parameters of the first coordination shell around Cu ions within the modified borate glasses (Cu–O); k = 2–9.6 (Å−1), R = 1–2.11 (Å), Hanning window
Full size table

In addition, Table 4 shows that the disorder parameter of Cu–O distance in the CuO phase is higher than that of Cu2O.

It was reported [36,37,38,39] that the structural coordination of CuO in various glass compositions was an elongated octahedron <CuO6>, which agrees well with our fitting results of the coordination number of Cu(II) in the modified borate glass for the first coordination shell.

Finally, the obtained findings shed light on the difference in the bioactivity of both borate and silicate matrices. It is known that borate glass is an amorphous/disorder material; the coexistence of CuO and Cu2O phases confirms the increase in the degree of disorder as extracted from EXAFS fitting. Modified borate glasses are known to be more efficient than that of their silicate partner (45S5 Hench Bioglass) resulting from the ease of conversion to their respective hydroxyapatite compartment and their low melting temperatures. The local structure of the matrices or the ease of degradation may be the result of such variation of the copper coordination within the glass matrix and reliability to be used in different medical applications, including hard tissue replacement [40,41,42,43].

4 Conclusions

This paper contributes to explaining the difference in the bioactivity of both borate and silicate matrices by following the short-range local structural change around the Cu atoms upon altering the Cu content (x), using XAFS. Near-edge XAFS (XANES) analysis confirms the formation of mixtures of Cu2O and CuO within the modified borate bioglass, while the silicate glass samples contain mainly Cu2O. According to the literature, a reasonable coexistence of both types of Cu ions has been proven to promote the formation of HCA, i.e., a suitable mixture of Cu(I) and (II) ions is needed for applications in bone bonding applications. Of our samples, the one with x = 4 seems the best. The XANES and EXAFS analyses yield consistent results regarding the relative amounts of Cu2O and CuO phases. Lastly, XAFS offers unique advantages for this type of functional materials analysis, including element specificity, short-range sensitivity, and the ability to explore local structures and oxidation states. As XANES/EXAFS analysis can deliver such chemical and structural information, we may conclude that these methods are well suited to shed light on the materials used in bone bonding applications.