2019 Volume 60 Issue 8 Pages 1707-1710
The 65ZnO–30P2O5–5Nb2O5 (ZnPNbO) glass is one of the promising candidate antibacterial biomaterials; it is a safe material since the release of Zn2+ ions from the glass is well-controlled. To discuss the origin of this controlled release from the view point of its glass structure, the amount and state of bridging oxygen in the glass was examined using X-ray photoelectron spectroscopy, and compared with those in 65CaO–30P2O5–5Nb2O5 (CaPNbO) glass. The number of P–O–P bonds in ZnPNbO glass was larger and that of P–O–Nb bonds was smaller than those in CaPNbO glass. This was linked with the formation of a large number of P–O–Zn bonds. The outermost shell electron density of the cations in ZnPNbO glass decreased, and as a result, the binding energy of each element increased. These might be closely related to the controlled release of Zn2+ ions.
Recently, Nb2O5-containing phosphate invert glasses are being investigated actively by our group to obtain new types of biomaterials with the ability to release effective therapeutic ions for inducing antibacterial properties and/or stimulating bone-forming ability.1–3) The glasses consist predominantly of the isolated phosphate groups and/or the shortest phosphate groups, i.e., ortho- and/or pyro-phosphates, which are denoted as QP0 and QP1, respectively. Bivalent cations, such as Ca2+, Mg2+, and Zn2+ ions connect the phosphate groups. These types of glasses are called “phosphate invert glasses”,4) denoted as PIGs, hereafter. The additive of Nb2O5 into PIGs play important roles in their glass formation and dissolution behavior.1,2) Sometimes, the phosphate groups have been reported to be cross-linked with a 4-oxygen-coordinated niobium group,5) assisting the formation of the glass network; with increasing Nb2O5 amount, the number of P–O–Nb bonds, which are stronger than the P–O–P bonds,6) increases. The formation of the P–O–Nb bonds controls the ion-release from the glasses in a Tris buffer solution.2)
Niobium is known as one of the elements with almost no toxic effect on our body (IC50 of niobium ions for MC3T3-E1 cells is 1.47 × 10−3 mol/L7,8)), and it enhances bone regeneration.8) The differentiation and mineralization of mouse-derived osteoblast-like cells, MC3T3-E1, were enhanced in Nb2O5-containing PIGs with the compositions of 60CaO–30P2O5–3Na2O–7Nb2O5 and 60CaO–30P2O5–7Na2O–3Nb2O5.9) Although the PIGs released Ca2+ and phosphate ions, most of the niobic ions would remain around the glass surface owing to low solubility, to form a gel layer.3) As a result, a trace amount of niobic ions would be released from the layer. The ion was believed to demonstrate a positive effect on bone regeneration.
When ZnO was included in CaO–P2O5–Nb2O5 invert glasses, antibacterial activity was reported; the PIGs with >35 mol% (effectively, >45%) of ZnO demonstrated antibacterial ability toward gram-positive and -negative bacteria.10) Exceedingly small amounts of Zn2+ ions (∼30 µM after 1 day) were released from the glasses (∼90 µM for 7 days). Typically, a 65ZnO–30P2O5–5Nb2O5 glass is one of the exceptional antibacterial materials.
The chemical durability of PIGs was improved by increasing the Ca/P ratio.2,11) 65CaO–30P2O5–5Nb2O5 glass exhibits high durability; however, the released amount of Ca2+ ions from the glass was 0.4 mM for 7 days, which was over 4 times larger than the amount of Zn2+ ions released from 65ZnO–30P2O5–Nb2O5 glass. ZnO is known as an intermediate oxide,12) i.e., in certain glasses, some of ZnO acts as a network former. In the phosphate glasses, Zn2+ ions coordinate with phosphate groups to form P–O–Zn bonds. As a result, the glass-forming ability and chemical durability of the glasses are improved.13)
We are interested in ZnO as an intermediate oxide that plays an important role in tuning the releasability of the bio-functional ions. Therefore, clarifying the structures of their Nb2O5-containing PIGs is considered to be essential.
The glass structures have been analyzed based on Raman and solid-state nuclear magnetic resonance spectroscopy results, to identify the existence of Qp0 and Qp1 groups and P–O–Zn.10) Further, we do not discuss their local structure, featuring a bridging oxygen (BO) and non-bridging oxygen (NBO). In the present work, the structural difference between 65CaO–30P2O5–5Nb2O5 and 65ZnO–30P2O5–5Nb2O5 was discussed based on X-ray photoelectron spectroscopy results, to make it useful for designing bio-functions.
Two types of PIGs with the nominal compositions of 65MO–30P2O5–5Nb2O5 (mol%, M = Ca, Zn) were prepared using ZnO (Kishida Chemicals, 99.5%), CaCO3 (Kishida Chemicals, 99.5%), Nb2O5 (Kishida Chemicals, 99.9%), and H3PO4 (Kishida Chemicals, 85%) as raw materials. Two compositions of M = Ca and Zn are denoted as CaPNbO and ZnPNbO, respectively. The reagents were mixed with distilled water in a beaker until slurry formation, and then the mixture was placed under an infrared lamp overnight. The dried mixture was melted in a platinum crucible at 1773 K for 1.8 ks, and then quenched to room temperature using an iron-pressing method.
The compositions of the resulting glasses were examined using energy dispersive X-ray fluorescence spectrometry (XRF; SEA2220A, SII Nano Technology Co.). The glass transition and crystallization temperatures, Tg and Tc, respectively, were determined using differential thermal analysis at a heating rate of 0.083 K/s (DTA; Thermoplus TG8120, Rigaku Co.). The index of process window IPW (i.e., the indicator of glass-forming ability) for each glass was calculated as follows: (Tc − Tg)/Tg [K/K].14)
X-ray photoelectron spectroscopy (XPS; PHI-5000 VersaProbe, ULVAC-PHI, Inc.) was performed for analyzing the local structure of the glasses. Aluminum was used as the target, and 1486.6 kV of voltage was applied. To eliminate the influence of the positive charge shift in the XPS measurement as much as possible, after cleaning the sample surface for 180 s using an Ar–ion gun, the XPS spectrum was collected. The binding energy was normalized with respect to the C 1s energy. 31P magic angle spinning nuclear magnetic resonance (MAS-NMR; JNM-ECA600 II, JEOL Ltd.) spectra were obtained to evaluate the glass structures at a Larmor frequency of 242.954 MHz in a 3.2 mm rotor spinning at 15 kHz under the conditions of a single pulse experiment with a 1.1 µs width, 5.02 s recycle delay, and cumulated number of 256 scans. Ammonium dihydrogen phosphate (NH4H2PO4, Kishida Chemical Co.) was used as a reference of 1 ppm.
The clear glasses could be prepared using quick iron-pressing. Table 1 shows the nominal and analyzed compositions of the glasses estimated from XRF and their Tg, Tc, and IPW from DTA. The amount of CaO or ZnO was analyzed to be as low as ∼5 mol%, from the nominal composition, while those of P2O5 and Nb2O5 were analyzed to be ∼5% and ∼1.5%, respectively. These differences may originate from the overlap in the characteristic X-ray peaks of P-Kα and Nb-Lα. In our experience, glasses with IPW > 0.1 can be easily prepared in small volume scales using a conventional melt-quenching method; ZnPNbO demonstrated satisfactory glass-forming ability, whereas relatively rapid quench was required for the glassification of CaPNbO (IPW = 0.07), since ZnO is an intermediate oxide with higher field strength than CaO (4.9 × 10−3 and 3.3 × 10−3 (valence/nm2) for ZnO and CaO,15) respectively).
Figure 1 shows 31P MAS-NMR spectra of the glasses. The spectra were fitted using Gaussian lines; both glasses consisted of ortho- and pyro-phosphates groups, i.e., QP0 and Qp1, respectively. CaPNbO glass constituted 29% of QP0 and 71% of QP1, while the ZnPNbO one constituted 17% of Qp0 and 83% of QP1. The substitution of Ca2+, coordination number 6 to 12,16) for Zn2+, coordination number 4 or 6,17) might affect the structure because of the phosphate chains shortage, with QP0 decreasing and QP1 increasing.
31P MAS-NMR spectra of CaPNbO and ZnPNbO glasses. Broken lines represent the considerable peaks deconvoluted using Gaussian lines.
There were almost no significant changes in the peak positions of QP1 and just a slight shift to the lower magnetic field side in that of QP0. The shift resulted owing to the coordination of Zn2+ with larger field strength around QP0. Further, there were differences in the full width at half maximum (FWHM) of their peaks (QP1 and QP0); the FWHM of the peaks in the spectrum of ZnPNbO glass was larger than those of CaPNbO glass. Certain different chemical states would exist around their structures. It should be considered that some amount of ZnO act as an intermediate oxide.
Figure 2 shows the O 1s XPS spectra of the glasses. The peak of ZnPNbO glass shifted to the higher binding energy side compared with that of CaPNbO one. The XPS spectra are reported to shift owing to changes in the binding energy of the inner shell electrons.18) When the electron density decreases, it becomes difficult for electrons to shield the positive charge of the nucleus, and the inner shell electrons approach the nucleus to shield its electrostatic force. Therefore, the bound energy of the inner shell electrons increases and shifts to the higher energy side. In the present glasses, the electronegativity of each atom is O: 3.44, Ca: 1.00, and Zn: 1.65. Owing to substitution of Ca2+ for Zn2+, the ionic bond property between oxygen and cation decreases and the electron density of oxygen decreases. As a result, it is considered that the spectrum of ZnPNbO glass shifted to the higher energy side when compared with the one from CaPNbO glass.
O 1s XPS spectra of CaPNbO and ZnPNbO glasses. Broken lines represent the considerable peaks deconvoluted using Gaussian lines.
Each peak was deconvoluted assuming Gaussian lines. Nb2O5 in phosphate invert glasses has been reported to act as an intermediate oxide.1,5,19,20) In the O 1s peaks of the glasses, they would include the information of non-bridging oxygen (NBO) at P–O− 21,22) and Nb–O− 23,24) bonds and that of bridging oxygen (BO) at P–O–P21,22) and P–O–Nb23,24) bonds. Moreover, as described above, for ZnPNbO glass, the presence of P–O–Zn bonds as a network former should be considered. It has been reported5) that for these glasses, Nb assumes the form of NbO4 and NbO6 and Zn is known to preferentially coordinate with phosphate groups when added in phosphate glasses. ZnO forms P–O–Zn bonds in aZnO-(1 − a) P2O5 (a = 0.50∼0.71, mol%) phosphate glasses, functioning as a part of the structure.
The percentage of each bond obtained by the deconvolution of the O 1s peaks is showed in Table 2. CaPNbO glass included totally 25% of BO consisting of P–O–P and P–O–Nb; the BO amount was small owing to the invert structure of the glass. Further, ZnPNbO glass included a large number of P–O–Zn bonds (25%) with a total of 20% P–O–P and P–O–Nb bonds. This result demonstrates that Zn, compared with Nb, has higher affinity to coordinate with phosphate groups. The increase in FWHMs of the Qp1 and Qp0 of the 31P MAS-NMR spectra shown in Fig. 1 would be affected by the coordination. In the present Nb2O5-containing PIGs, it was observed that not only Ca2+ ion coordinated to the NBO of P–O− was simply substituted with Zn2+ ion, but also a large amount of Zn2+ ion forms BO to form P–O–Zn bond, contributing to glass network formation. The formation of the P–O–Zn bond increased the number of P–O–P bonds (QP1) and reduced the number of P–O–Nb bonds. As a result, the increase in BO amount in the ZnPNbO glass induced its satisfactory forming ability (IPW ∼ 0.13).
The network connectivity (NC) of phosphate glasses is estimated by weighted averages of QPn (n = 1∼3), except n = 0, which forms no network.24,25) From the result of Fig. 1, the NCs of CaPNbO and ZnPNbO are estimated as 0.71 and 0.83, respectively. Greater NC leads to an increase in bond strength of the entire glass.26) When the electron density decreases, it becomes difficult for the electrons to shield the positive charge of the nucleus. As a result, the inner shell electrons approach the nucleus and try to shield the electrostatic force: the binding energy of the inner shell electron increases.27) That is, when the ion bonding property between oxygen and cation decreases (i.e., the covalent bonding property increases), the energy density of oxygen decreases and the peak of the binding energy shifts to the higher energy side. Therefore, the increase in XPS binding energy is considered to be owing to an increase in bond strength of the entire glass.
Figure 3 shows P 2p and Nb 3d spectra. The origins of their binding energies were referred to Refs. 22 and 23, respectively. In the case of ZnPNbO glass, the binding energy owing to the inner shell electrons shifted to the higher energy side. Thus, it is considered that the electron densities of P and Nb decreased, as well as that of O. The present PIGs contain a large amount of NBO: numerous cations are coordinated to NBO. It has been reported27) that in this type of glass, the inner shell electrons of the cations behave as though they are shielded with the Coulombic field generated by oxygen. Thus, the outer shell electrons of P and Nb have limited effect on their own Coulombic field, and the electronic state of oxygen dominates the binding energy of the inner shell electron in the cation. Moreover, since the outermost shell electron density of NBO is higher than that of BO, the cation coordinated with more NBO has higher outermost shell electron density.18) Therefore, in the case of ZnPNbO glass, the outermost shell electron densities of the cations decrease, and as a result, the XPS peaks of the glass are considered to shift to the higher energy side; this might be related to the controlled release of Zn2+ ions.
P 2p and Nb 3d XPS spectra of CaPNbO and ZnPNbO glasses.
The release of antibacterial Zn2+ ions were also effectively controlled, despite high content of ZnO.10) ZnPNbO glass was observed to include 25% of P–O–Zn bond. This ions-release control would be explained based on the formation of the bond, which acts as a glass network former.
CaPNbO and ZnPNbO glasses consisted of the shortest phosphate chains (QP1) and isolated phosphate groups (QP0), with ionically connected bivalent cations (Ca2+ or Zn2+). Some Nb2O5 acted as a glass network former. In ZnPNbO glass, a large number of P–O–Zn bonds formed dominantly, thus increasing the number of P–O–P bonds (QP1) and reducing the number of P–O–Nb bonds; ZnO played an important role in the easy formation of the glass. The XPS spectrum of each element in ZnPNbO glass shifted to the higher binding energy side compared with that in CaPNbO glass. This was considered to originate from the decrease in the outermost shell electron densities of cations.
Our previous work reported that ZnPNbO glass demonstrated exceptional antibacterial activity and high chemical durability; these are significant properties for designing biomaterials. The present work demonstrated that through discussion regarding the local structure of the glass using XPS analysis, the release control of an antibacterial ion would originate from some of the ZnO in the glass that acts as a glass network former.
This work was supported in part by JSPS KAKENHI Grant Number 26289238 and NITech Research Promotion Grants.
