Introduction

Numerous studies have confirmed that the phosphate glasses and glass-ceramics derived from the base CaO–P2O5 system meet the main criteria to be used as bioactive materials. Due to appropriate dissolution properties and the chemical composition which is similar to that of the mineral phase of the bone, the phosphate glasses in the system P2O5–CaO–MgO–Na2O became promising materials in bone tissue engineering [1–5]. This glass system allows variation of properties such as solubility or crystallization tendency by only minor chemical composition adjustments.

Phosphorous pentoxide belongs to the network-forming oxides which can form single-component glasses. The structure of phosphate glasses is specific, and consequently the behavior of these glasses differs in regard to the most widely used silicate glasses. The basic building units of phosphate glasses are the P-tetrahedra. These tetrahedra link through covalent bridging oxygens. Depending on the number of bridging oxygens per tetrahedron, the tetrahedra are classified using the Q i terminology [i represents the number of bridging oxygens (BO) per tetrahedron]. The structure of P2O5 glasses consists of only Q 3 phosphate tetrahedral units that form a three-dimensional network. The addition of modifying metal oxides results in depolymerization of this network breaking the P–O–P bonds with the creation of non-bridging oxygens (NBO) and changing from a cross-linked Q 3 network to chain-like Q 2 structures and further to depolymerized Q 1/Q 0 glasses with increasing [O]/[P] ratio. Acidic phosphate-based glasses (ultraphosphate glasses) with >50 mol% P2O5 have the structure of a three-dimensional network consisting of Q 3 and Q 2 groups and are of low chemical durability. By increasing the content of modifying oxide (x), the polyphosphate glasses where x > 0.5 and [O]/[P] > 3 can be obtained. Their structure consists of chains formed by Q 2 groups terminated with Q 1 groups. The average chain length decreases with increasing [O]/[P] ratio. At the [O]/[P] = 3.5 (pyrophosphate stoichiometry, x = 0.67), the structure is dominated by phosphate dimers, two Q 1 groups connected by a bridging oxygen. At x = 0.75, only isolated orthophosphate groups Q 0 are present. These glasses degrade more slowly in different aqueous solutions than ones with P2O5 > 50 mol% due to the reduction in Q 2 units in the structure [6–8]. By changing the composition, the glass degradation rate as well as its bioactivity can be tailored. Additionally, to improve biological response of the phosphate glass bone implants, trace elements with therapeutic effects (Zn, Li, Sr, Cu, Fe, Mn, etc.) are often added in the base glass composition [9–14]. These glasses can be transformed successfully into biodegradable glass-ceramics by crystallization process. The bioactivity behavior of this material is strongly connected with the presence of calcium-phosphate crystalline phases in the glass matrix [3, 15, 16]. To obtain glass-ceramics material, the glass powder sintering route is frequently employed [15–20]. In such a way, the porous 3D-bioactive glass and glass-ceramics scaffold for bone tissue engineering can be fabricated [21, 22]. Therefore, in order to optimize the parameters for preparation of these bioactive materials, it is necessary to examine the crystallization and sintering behaviors of the selected bioactive glass composition [23–26].

In this study, the crystallization and sintering behaviors of the polyphosphate P2O5–CaO–MgO–Na2O–SiO2 glass doped with trace elements—Zn and Mn—were investigated. The glass was prepared by standard melt-quenching technique, and the heating experiments on bulk and powder glass samples were performed. The crystallization and sintering processes were analyzed using differential thermal analysis (DTA) and hot-stage microscopy (HSM). Phase composition and microstructure of as-prepared glass-ceramics were analyzed by XRD and SEM methods. The structural changes have been analyzed by Fourier transform infrared spectroscopy (FTIR).

Experimental

The glass of the selected composition was obtained by melting a homogenous mixture of: reagent-grade (NH4)2HPO4, K2CO3, CaCO3, MgO, SiO2, ZnO and MnO. The melting was performed in porcelain crucible at 1100 °C for 1 h, and the melt was cast on steel plate. XRD analysis confirmed that the glassy material is obtained. The sample was transparent without residual gas bubbles.

The crystallization behavior of the glass under non-isothermal conditions was analyzed using a differential thermal analysis (DTA)—Netzsch STA 409 EP device with Al2O3 powder as the reference material. Powdered sample (<0.048 mm) was heated from 20 to 900 °C at a heating rate of 10 °C min−1. To determine the activation energy of crystallization, the DTA crystallization peak temperatures T p were recorded at heating rates of 5–20 °C min−1.

The isothermal crystallization of bulk glass samples was performed by heating the samples in an electric furnace Carbolite CWF 13/13 with automatic regulation and temperature accuracy of ±1 °C, up to the crystallization temperatures previously determined by DTA, at a heating rate of 5 °C min−1 and maintained at these temperatures for different times. After heat treatment, the samples were removed from the furnace, cooled in air and prepared for analyses.

For determining the sintering behavior of the glass powder samples, a hot-stage microscope (HSM), E. Leitz Wetzlar, supplied with Canon camera was used. Glass powder samples (<0.048 mm) were hand-pressed into cylinders, and the specimens were placed on a platinum plate and then on alumina support which was in contact with thermocouple. The temperature was measured with (Pt/Rh/Pt) thermocouple at a heating rate of 5 °C min−1. The sample-shape images were analyzed using computer software for image analyzing.

The samples for sintering were prepared using the glass powder (<0.048 mm) cold-pressed in Manfredi C 95 laboratory hydraulic press at 30 MPa. The samples in a form of pellets (Ø 10 mm) were fabricated. The temperatures of sintering were selected according to DTA and HSM experiments. The samples were heated at a rate of 5 °C min−1 up to the chosen temperatures and then maintained at these temperatures for various times. After cooling, the sintered samples were crushed in an agate mortar and the favorable fractions were selected for SEM analysis. The microstructure of the crystallized glass samples was determined using a MIRA3 XMU TESCAN microscope. The samples were gold-coated in a Leica SCD005 device.

The phase composition of the crystallized bulk glass samples and sintered glass-ceramics was examined by XRD method using a Philips PW-1710 automated diffractometer with Cu tube operated at 40 kV and 30 mA. The instrument was equipped with a diffracted beam curved graphite monochromator and Xe-filled proportional counter. The diffraction data were collected in the 2θ Bragg angle from 5° to 70°, counting for 1 s.

The structural characteristics of the sintered glass-ceramics sample were studied by Fourier transform infrared spectroscopy (FTIR). All IR absorption spectra were recorded in the wave number range 400–1600 cm−1 (transmission mode) using a Thermo Scientific Nicolet iS50 spectrometer. The resolution used was 4 cm−1, and the number of scans was 32.

Results and discussion

Chemical composition and the structure of glass

The nominal and analyzed chemical compositions of the glass are given in Table 1.

Table 1 Chemical compositions of the glass
Full size table

As given in Table 1, the parent glass belongs to the polyphosphate glasses (x > 0.5), where x is the molar fraction of the network modifier oxides. The network of these glasses is composed of Q 2 chains terminated by Q 1 tetrahedra (one bridging and three non-bridging oxygens per tetrahedron). For analysis of glass structure, we considered this glass as a binary one: x Me2/ν O·(1 − x)NFO, where Me is modifier ion, ν is its valence, NF are network former ions, and x is the molar fraction. In this case, [Me2/ν O] = ([M2O] + [MO]) = [K2O] + [CaO] + [MgO] + [ZnO] + [MnO] and [NFO] = [P2O5] + [SiO2]. Based on its composition, glass was treated as a binary 0.52 Me2/ν O·0.48NFO. For binary phosphate glasses with the molar fraction x of 0.5 < x < 0.67, the fraction of Q 1 and Q 2 tetrahedra are given as f(Q 1) = (2x − 1)/(1 − x) and f(Q 2) = (2 − 3x)/(1 − x) and the average tetrahedron chain length (L av) is given as L av = 2(1 − x)/(2x − 1) [6]. Accordingly, f(Q 1) = 0.08, f(Q 2) = 0.92 and L av = 24 were calculated. Regarding the values obtained, it may be concluded that the structure of this glass consists of phosphate chains containing 24 tetrahedra, where 22 of them contain two bridging oxygens and two terminal tetrahedra with one bridging oxygen. It can be considered that the modifier ions are placed in the cavities between tangled phosphate chains.

Thermal analysis (DTA)

Figure 1 shows the DTA curve of powder glass sample (<0.048 mm) recorded at a heating rate of 10 °C min−1. The DTA curve (Fig. 1) shows an exothermal peak at T p = 473 °C, indicating glass crystallization. The glass liquidus temperature T l = 585 °C is determined as a peak of the small melting endotherm. Also, the glass transition temperature T g = 305 °C and onset crystallization temperature T x = 390 °C can be determined on DTA curve. The glass stability (GS) parameter, ΔT = T x − T g = 85 °C, indicated a low tendency toward crystallization of this glass. This parameter strongly depends on glass composition, and generally, the higher is the value of ΔT for a certain glass, the higher is its stability against crystallization. For polyphosphate glass compositions, the chain length of the phosphate units affects markedly on glass stability. The increase in glass stability is caused by a lower mobility of long phosphate units in the melt [6, 8].

Fig. 1
figure 1

DTA curve recorded at a heating rate of 10 °C min−1 for sample particle size <0.048 mm

Full size image

Isothermal crystallization of the bulk glass samples

The bulk glass samples were isothermally treated at temperature T c = 500 °C for t = 1, 3, 6, 12, 24 and 72 h. It was detected that for times shorter than 12 h, the glass does not crystallize. The XRD patterns collected for these samples exhibited the shapes which are distinctive for an amorphous material, Fig. 2.

Fig. 2
figure 2

XRD patterns of the bulk glass sample isothermally heated at T c = 500 °C for t = 1, 3, 6, 12, 24 and 72 h

Full size image

At heating time t = 12 h, the glass sample commences to crystallize and crystalline phases determined are: KPO3 (JCPDS - 74 - 0511) and Ca2 K P3O10 (JCPDS - 40 - 0190). With increasing the heating time up to t = 24 h, the formation of the new crystalline phase KCaPO4 (JCPDS - 33 -1002) was determined. When the heating time is increased to t = 72 h, an additional change of phase composition including formation of the new phases: α-Ca2P2O7 (JCPDS - 45-1061), γ-Ca2P2O7 (JCPDS - 23-0871), K3P3O9 (JCPDS - 70 - 0048) and K4P2O4 (JCPDS - 20 - 0921), is registered. The crystal grow rates are so low that after 72 h, a large quantity of glassy phase is still present in the sample, Fig. 2.

The SEM micrographs of the fractured surface of the bulk glass samples heated at T c = 500 °C revealed that the process of glass/crystal transformation was carried out by simultaneous acting of surface and volume mechanism of crystallization, Fig. 3. In Fig. 3a, b, the surface-crystallized layer with needle-like dendritic morphology of the crystal growth is clearly visible. Under higher magnification, numerous crystallites in the glass matrix can be observed, Fig. 3c, d.

Fig. 3
figure 3

SEM micrograph of the fractured surface of the bulk glass samples heated at T c = 500 °C for t = 12 h: a crystallized surface layer; b crystal/glass interface at higher magnification; c, d crystallites in glass matrix

Full size image

Glass powder sintering

In Fig. 4, the photomicrograph of the glass sample shapes collected on a hot-stage microscope (HSM) is shown. The characteristic temperatures: first shrinkage (T FS), maximum shrinkage (T MS), deformation (T D), sphere (T S), half ball (T HB) and flow (T F), were determined using the changes of relative area A/A o as a function of temperature, where A o is initial area and A is area at the temperature T.

Fig. 4
figure 4

Photomicrograph of the glass sample shapes collected on hot-stage microscopy (HSM)

Full size image

The dependency of A/A o on temperature is shown in Fig. 5. The glass transition temperature T g and onset crystallization temperature T x of the glass previously determined by DTA measurements (Fig. 1) are marked in the figure. The temperatures which represent glass melting (T HB and T F) determined by HSM (Fig. 5) are slightly lower than the DTA liquidus temperature (T l). It should be noted that the melting peak on DTA curve (Fig. 1) is not well expressed and T l could not be determined with high accuracy.

Fig. 5
figure 5

A/A o versus temperature of glass sample obtained from HSM

Full size image

According to the shrinkage curve obtained, Fig. 5, the maximum shrinkage of the sample (A/A o = 0.70) was determined at T = 385 °C. Further heating up to T = 540 °C induces the expansion of sample (A/A o = 0.86) and after that an abrupt decrease in A/A o ratio due to the melting of the sample.

To determine that the sintering and crystallization processes proceed independently, it is useful to compare the results of HSM and DTA obtained in the same temperature range and heating conditions. As proposed by Lara et al. [20], the competition between glass sintering and crystallization of glass powders during heating can be described using a parameter S C = T x − T MS, where T x is onset crystallization temperature and T MS is maximum shrinkage temperature. This parameter is a measure of the independency of sintering and crystallization processes. During sintering of glass powder compact, the surface crystallization of glass particles as a concomitant process can occur, and in such a way, the sintering kinetics is slowing down.

T x > T MS for this glass indicates that the sintering process precedes the crystallization of glass particles. Therefore, a full densification of glass powder sample can be attained before the crystallization starts.

To obtain glass-ceramics, the glass powder compact was sintered at T c = 500 °C for t = 12 h. In Fig. 6, the results of XRD analysis of sintered sample are shown. As shown in Fig. 6, the phase composition of glass-ceramics prepared via sintering and crystallization of glass powder compact is complex and the crystalline phases formed are: KPO3 (JCPDS - 74-0511), K2Mg(PO3)4 (JCPDS - 24 - 0877), Ca(PO3)2 (JCPDS - 79-0700) and KCa(PO3)3 (JCPDS - 39-1408).

Fig. 6
figure 6

XRD pattern of sintered glass powder compact

Full size image

In comparison with the crystallization of bulk glass sample, a more intensive surface crystallization of glass particles took place during sintering. The higher specific surface of fine glass powder enhanced the surface nucleation. As may be seen, the phase compositions of crystallized bulk and powder glass samples are different. Only KPO3 phase was determined in both samples, Figs. 2 and 6.

The structural transformations which appeared during heating of glass powder compact at T c = 500 °C for t = 12 h are visible on the FTIR spectra, Fig. 7.

Fig. 7
figure 7

FTIR spectra of a as-quenched glass and b sintered glass powder compact at T = 500 °C for t = 12 h

Full size image

As shown in Fig. 7, the phosphate as-quenched glass revealed the characteristic bands in IR spectra in the wave number range 400–1400 cm−1 and the peaks assigned are: ~485, 715, 875, 1090 and 1260. Based on previous FTIR studies of the structure of phosphate glasses, the peaks at 485 cm−1 can be assigned to bending vibration of bridging phosphorous δ (P–O–P). The peak at 715 cm−1 belongs to P–O–P symmetric stretching vibration, ν s (P–O–P), of the bridging oxygen atoms bonded to a phosphorus atom in a Q 2 phosphate tetrahedron, and one at 875 cm−1 is due to the asymmetric vibration of P–O–P bond, ν as (P–O–P). At 1090 and 1260 cm−1, the peaks which belong to asymmetric stretching of (PO3)− groups (characteristic of Q 1 structural units) and the asymmetric stretching mode of (O–P–O), ν as (PO2) (characteristic of Q 2 structural units), are visible. The crystallization of glass was expressed on FTIR curves by transformation of the characteristic bands into several new peaks. These new peaks appeared in wave number ranges which correspond to the vibration modes of metaphosphate (PO3)− groups [3, 27–30]. These results are in accordance with functional groups of the crystalline phases precipitated in glass-ceramics, Fig. 6.

The glass-ceramics prepared by sintering of the glass powder compact at T c = 500 °C for t = 12 h has been exposed to SEM analysis, and the selected micrographs of free and fractured sample surface are shown in Fig. 8.

Fig. 8
figure 8

SEM micrographs of glass-ceramics sample obtained by glass powder sintering at T = 500 °C for t = 12 h: a, b free surface and c, d fractured surface

Full size image

In Fig. 8a, b, the interconnected glass particles with numerous crystallites on the surface are visible. The pores appeared as the small black spots between the glass particles. The micrographs of the fractured sample surface, Fig. 8c, d, revealed a microstructure which has been formed by surface growth of crystals with a plate-like morphology. Also, in Fig. 8c, d, the presence of the closed pores in a sintered compact is visible. It can be considered that these pores were induced by the crystallization of the glass particles.

Non-isothermal crystallization kinetics of glass

Among several analytical methods available for analysis of the DTA data and determination of the activation energy for crystallization, the Kissinger method is the most frequently used [31]:

$$ \ln \frac{\beta }{{T_{\text{p}}^{2} }} = - \frac{{E_{\text{a}} }}{{R \times T_{\text{p}} }} + {\text{const}}. $$
(1)

where E a—activation energy for crystallization, β—heating rate, T p—DTA crystallization peak temperature, and R—gas constant.

As reported by Matusita and Sakka [32], Eq. (1) is valid only when crystal growth occurs on a fixed number of nuclei and the surface crystallization of glass dominates. Accordingly, the glass powder of the smallest particle size <0.048 mm was heated at heating rates β of 5, 10, 15 and 20 °C min−1. The activation energy of crystallization E a was determined from the slope of the line of the Kissinger plot ln(β n/T 2p ) versus 1/T p. In Fig. 9, the DTA crystallization peaks of glass recorded at different heating rates are shown, and the values of T p are listed in Table 2. The Kissinger plot ln (β/T 2p ) versus 1/T p is shown in Fig. 10.

Fig. 9
figure 9

DTA crystallization peaks of glass recorded at different heating rates

Full size image
Table 2 Temperature of crystallization peaks, T p, for glass powder sample heated at different heating rates β
Full size table
Fig. 10
figure 10

Kissinger plot ln(β/T 2p ) versus 1/T p

Full size image

An activation energy E of crystal growth was calculated from the slope of the line, E a,k = (246.36 ± 19.74) kJ mol−1.

The method proposed by Ozawa can be also applied for calculation of the activation energy [33]:

$$ \ln (\beta ) = - (E_{\text{oz}} /RT_{\text{p}} ) + {\text{const}}. $$
(2)

The activation energy of crystallization E a,oz = (258.80 ± 19.70) kJ mol−1 was determined from the slope of the line of the plot ln(β) versus 1/T p, Fig. 11. This value is in good agreement with Kissinger method.

Fig. 11
figure 11

Ozawa plot ln(β) versus 1/T p

Full size image

Conclusions

The subject of these investigations was the crystallization and sintering behaviors of 45 P2O5·15CaO·10MgO·25K2O·3SiO2·ZnO·MnO (mol%) glass prepared by standard melt-quenching technique. Low crystallization ability and a complex crystallization process which took place under simultaneous acting of the surface and volume mechanisms of crystallization were determined for this glass. The characteristic dendritic morphology of crystal growth at the surface has been observed for bulk glass. Comparing the results of DTA and HSM analyses, it was shown that the sintering and crystallization processes proceed independently. The phosphate glass-ceramics was fabricated by sintering and crystallization of glass powder compact at T c = 500 °C for t = 12 h. Surface crystallization of glass particles and a plate-like morphology of crystals growth were determined. The glass-ceramics with complex phase composition was obtained, and the crystalline phases determined are: KPO3, K2Mg(PO3)4, Ca(PO3)2 and KCa(PO3)3. The activation energies of crystal growth calculated by Kissinger and Ozawa relations are in good agreement.