# Fileset

[133_24096.pdf](https://mdr.nims.go.jp/filesets/beba1168-a934-4997-b0ce-dc582fe30051/download)

## Creator

Hirokazu Masai, Takahiro Ohkubo, [Yohei Onodera](https://orcid.org/0000-0002-3080-6991), Yasuhiro Fujii, Seiya Shimono, Akitoshi Koreeda

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Effect of borate substitution on zinc phosphate glasses](https://mdr.nims.go.jp/datasets/f5123d7a-30c3-4665-a159-e6f6fb5f6894)

## Fulltext

Effect of borate substitution on zinc phosphate glassesJCS-Japan第133巻　第 1 号　2025年 1 月 1 日発行（毎月1回1日発行）　ISSN 1348-6535 CODEN: JCSJEWJanuary2025vol.133Journal of the Ceramic Society of JapanFULL PAPEREffect of borate substitution on zinc phosphate glassesHirokazu Masai1,³, Takahiro Ohkubo2, Yohei Onodera3, Yasuhiro Fujii4,5,Seiya Shimono6 and Akitoshi Koreeda71National Institute of Advanced Industrial Science and Technology, 1–8–31 Midorigaoka, Ikeda, Osaka 563–8577, Japan2Graduate School & Faculty of Engineering, Chiba University, 1–33 Yayoi-cho, Chiba 263–8522, Japan3Center for Basic Research on Materials, National Institute for Materials Science, 1–2–1 Sengen, Tsukuba, Ibaraki 305–0047, Japan4Institute for Open and Transdisciplinary Research Initiatives, Osaka University, 2–1 Yamada-Oka, Suita, Osaka 565–0871, Japan5Research Organization of Science and Technology, Ritsumeikan University, 1–1–1 Noji-higashi, Kusatsu, Shiga 525–8577, Japan6Japan Synchrotron Radiation Research Institute (JASRI), 1–1–1 Kouto, Sayo-cho, Sayo-gun, Hyogo 679–5198, Japan7Department of Physical Sciences, Ritsumeikan University, 1–1–1 Nojihigashi, Kusatsu, Shiga 525–8577, JapanThe combination of glass-forming oxides can induce phase separation at the atomistic level, particularly when amixture involves both ionic and covalent bonds. In this study, we report physical properties of B2O3-substitutedZnO–P2O5 (ZP) glass and the substitution effect of B2O3 on the structure of ZP glasses via magic angle spinningnuclear magnetic resonance (MAS NMR), X-ray diffraction, small angle X-ray scattering, positron annihilationspectroscopy and inelastic light scattering measurements. The 11B and 31P MAS NMR results suggest that thecoordination states of boron are affected by the number of bridging oxygen atoms in the substituted PO4 chainstructure of the ZP glass. At 5mol% or less B2O3 substitution, an emerging of atomistic phase separation is notconfirmed. On the contrary, at 10mol% B2O3 substitution, three-coordinated boron formation is confirmed in11B MAS NMR in addition to a correlation between the BO4/2 units. The substitution of a large amount of B2O3confirmed the local coordination change of zinc cations in addition to the formation of a cluster-like structure inthe low-Q region via small angle X-ray scattering.Key-words : Glass, Phosphate, Borate, Structure, NMR, Diffraction, Inelastic light scattering[Received August 31, 2024; Accepted October 13, 2024; Published online November 20, 2024]1. IntroductionThe metaloxane network connection in oxide glass isimportant not only for the vitrification of monolithic bulkglass without crystallization but also for the physical prop-erties originating from the connectivity of constituting cat-ion unit.1) Based on its glass-forming ability, SiO2, GeO2,B2O3, and P2O5 are classified as network former (NWF)units.2) However, because of the ionic P=O bond, P2O5is different from other NWF units, and the structure ofphosphate glass varies greatly depending on the chemicalcomposition.3–13) The lower covalency of phosphate glassresults in lower phonon vibrational energies comparedwith those of conventional silicate and borate glasses,which is advantageous for luminescence applications. Inaddition to lower phonon vibrational energy, the ionicphosphate units have a high affinity for active cations,allowing for a more uniform distribution and higher dop-ing of active cations compared to conventional SiO2 orB2O3 glasses. Therefore, phosphate glasses are sometimesused as a host material for activators.14–18)It has been reported that the connectivity of the con-stituents of glassy materials is not homogeneous from anatomistic point of view, even in pristine SiO2 glass.19) Ifthere are different metaloxane units exhibiting differentcovalent/ionic bonds, they tend to phase separate, at leastatomistic level. The phase separation is a kind of thermo-dynamically stabilization processes for liquids, and it gen-erally accelerates crystallization of glass. The conventionalspinodal phase separation is often observed in multi-component glass where ionic and covalently bonded unitscoexists. The phase separation is a density fluctuation ofcomponents of a glass, often leading to a loss of opticaltransparency.It was reported that 60ZnO–(40 ¹ x)P2O5–xB2O3 glassexhibits macroscopic phase separation or crystallization atthe intermediate chemical compositions.18) In the paper,Sn2+ was used as an activator, whose luminescent proper-ties depend on the chemical composition of glass, i.e. frac-tions of P2O5 and B2O3. Since Sn2+ possesses the electronsin the outermost shell in both ground and excitation states,the luminescence is strongly affected by the surroundingregion. In other words, the luminescence of Sn2+ is con-sidered to be an indicator for the local structure of glass.However, detailed structures, such as the structural rela-³ Corresponding author: H. Masai; E-mail: hirokazu.masai@aist.go.jpJournal of the Ceramic Society of Japan 133 [1] 1-8 2025DOI https://doi.org/10.2109/jcersj2.24096 JCS-Japan©2024 The Ceramic Society of Japan 1This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/),which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.https://doi.org/10.2109/jcersj2.24096https://creativecommons.org/licenses/by/4.0/tionships among the phosphate and borate units, have notbeen fully clarified. For example, it is not clear whetheratomistic phase separation20,21) (not a conventional macro-scopic phase separation observable by naked eyes) occursin ZnO–P2O5–B2O3 glasses containing small fraction ofB2O3.Recently, we reported the three-dimensional (3-D) struc-ture of binary zinc phosphate (ZP) glasses consisting ofZnO (intermediate group) and phosphate units constructedby reverse Monte Carlo modeling based on 31P nuclearmagnetic resonance (NMR), neutron and X-ray diffraction,and Zn K-edge X-ray absorption fine structure (XAFS).22)Based on various data obtained by different measurementtechniques, we find that the main phosphate network wassubstituted by an intermediate ZnO subnetwork, depend-ing on the chemical composition. Therefore, combinatorialcharacterization is important to investigate the structure ofglasses.22–26)Here, we examined B2O3-substituted ZP glasses and thecomposition-dependent structures using a combination ofinelastic light scattering, positron annihilation spectros-copy, solid-state magic angle spinning (MAS) NMR,X-ray diffraction, and small angle X-ray scattering. Wefocused on the atomistic phase separation behavior ofB2O3-substituted ZP glasses and the correlation of dataobtained using different measurement techniques.2. Experimental2.1 Preparation of B2O3-substituted ternaryzinc phosphate glassesTernary 60ZnO–(40 ¹ x)P2O5–xB2O3 (ZPBx) glasses(x = 0, 1, 2.5, 5, and 10) were prepared via a conventionalmelt-quenching method using a platinum crucible.27)Batches comprising ZnO (99.99%), (NH4)2HPO4 (99%),and B2O3 (99.9%) were initially calcined at 800 °C for 3 hin an ambient atmosphere. The calcined solid was thenmelted at 1,100 °C for 30min in the ambient atmosphere.The glass melt was quenched on a stainless-steel platemaintained at approximately 200 °C and then annealed atthe glass transition temperature Tg for 1 h. The sampleswere mechanically polished to obtain mirror surfaces. Thesample sizes were approximately 10mm © 10mm © 1mmfor the optical measurements.2.2 Analysis methodsTg was determined using a differential thermal analysis(DTA) system operating at a heating rate of 10 °C/minusing a TG8120 (Rigaku, Japan). The densities were mea-sured using the Archimedes method with water at the roomtemperature. We measured the refractive indices of thesamples using a prism coupler with 473, 633, 1,319, and1,553 nm light sources (Metricon, N.J., U.S.A.). The errorin the measurement was 10¹4.The Brillouin shifts ¯B of the glasses were measuredusing the high-resolution modification of a Sandercock FPsystem.28) The excitation laser was a frequency-doubleddiode-pumped solid-state (DPSS) Nd: yttrium-aluminum-garnet laser oscillating in a single longitudinal mode at532 nm (Oxxius SLIM-532 300mW). In addition to theexcitation laser source, a second weak reference laser wasemployed as an independent reference signal to stabilizethe FP. The reference laser was a small DPSS Nd: yttrium-vanadate (Nd:YVO4) laser module (Photonic Products300-0088-01, 4mW) oscillating in a single transversemode (TEM00), with two to three longitudinal modesseparated by 120GHz. The longitudinal sound velocitiesVL listed in Table 1 were calculated using the relationVL = ¯B­/2n532, where ¯B, ­, and n532 are the Brillouinshift, wavelength of the incident light (= 532 nm), andrefractive index at 532 nm, respectively. The n532 valueswere calculated using the Cauchy relation with the refrac-tive indices at different wavelengths.Positron annihilation lifetime measurements were per-formed using a PSA TypeL-II (Toyo Seiko Co., Ltd.). The22Na source, which was encapsulated with Kaptonμ film,was used for measurement.29) The accumulated counts foreach sample was 107.31P and 11B MAS NMR spectra were obtained to exam-ine the local coordination states of each metal oxide. The31P MAS NMR spectra of the glasses were acquired on aDELTA 600 spectrometer (JEOL, Japan) at a frequency of242.95MHz with 15 kHz spinning rate and 500 s relaxa-tion delay for 16 scans. The chemical shifts were estimatedwith respect to an 85% H3PO4 aqueous solution (0 ppm).The 11B MAS NMR spectra of the glasses were acquiredat 192.6MHz with 15 kHz spinning rate. For each sample,256 acquisitions were obtained with a pulse delay of 3 sand a pulse width of 0.3¯s with a tip angle of 15°. The11B MAS spectra were corrected and referenced against a1M H3BO3 aqueous solution at 19.6 ppm. Spectral de-convolution was performed using the dmfit 2002 programwith a “Q-mas 1/2” model that includes the assumption ofthe three boron species to estimate the population andNMR parameters of the boron species. The chemical shiftTable 1. Physical parameters of the ZPBx glass. The sound velocities were calculated using values of theBrillouin shift ¯B and refractive index of incident light (532 nm)B2O3 fraction(mol%)IDTg (°C)(«3 °C)Molar volume(cm3mol¹1)Refractive index at532 nm («0.0003)Longitudinal sound velocity(m s¹1) («1m s¹1)C11 (GPa)0 ZP 421 32.39 1.5792 4,563 66.591.0 ZPB1 420 31.95 1.5821 4,570 68.562.5 ZPB2.5 422 30.97 1.5906 4,686 73.595.0 ZPB5 426 30.04 1.6040 4,869 80.5210.0 ZPB10 475 28.02 1.6213 5,112 91.78Masai et al.: Effect of borate substitution on zinc phosphate glassesJCS-Japan2of 11B is influenced by its first coordination numbers, BO3and BO4. It is necessary to introduce two three-coordinatedboron species for ring (B3O3 two three-coordinated boronspecies for ring and non-ring structures) and non-ringstructures and four-coordinated boron (BO4).A high-energy X-ray diffraction experiment was per-formed at the BL04B2 beamline at the SPring-8 synchro-tron radiation facility (Hyogo, Japan) using a two-axisdiffractometer dedicated to the study of the disorderedmaterials.30) The incident X-ray energy was 61.43 keV(x = 0, 1, 2.5, and 5) or 112.79 keV (x = 10). The raw datawere corrected for polarization, absorption, and back-ground, and the contribution of Compton scattering wassubtracted using a standard data analysis software.30)Small-angle X-ray scattering spectra were measured atthe BL19B2 beamline at SPring-8. The incident X-rayenergy was 30 keV, and the exposure duration was 90 s.The camera distance was 3045mm.The Zn K-edge (9.66 keV) XAFS spectra were mea-sured using the BL01B1 beamline at SPring-8. The mea-surements were performed using a Si(111) double-crystalmonochromator in transmission mode (Quick Scan meth-od) at RT. Pellet samples for the measurements were pre-pared by mixing a granular sample with boron nitride. Thecorresponding analyses were performed using the Athenasoftware.31)3. Results and discussionThe chemical composition of the glass was 60ZnO–(40 ¹ x)P2O5–xB2O3 (ZPBx). Since the B2O3 fraction isused to replace the P2O5 fraction, we refer here as B2O3-poor (x = 1, 2.5, and 5) and B2O3-rich (x = 10) glasses.Note that the total number of cations is not changed by theB2O3 substitution. The values of Tg increased with increas-ing B2O3 fraction, suggesting that a stronger network wasformed by the substitution of B2O3 units. Recently, wehave reported that cation-based composition is suitable fordiscussion of phosphate-based low-melting glasses.32)Although the report suggests that Tg can be approximatedto some extent by linear regression,32) the current Tg valuesare not proportional to the B2O3 fraction, especially incompositions with low B2O3 fraction. It suggests that Tg ofB2O3-poor compositions is dominated by phosphate units.Next, we examine several macroscopic properties of ZPBxglasses using inelastic light-scattering approaches. Thevalues of Tg, molar volume, refractive index, and longi-tudinal sound velocity VL are summarized in Table 1, inaddition to the longitudinal elastic modulus c11 valuescalculated from the density, refractive index, and longi-tudinal sound velocity of the ZPBx glasses. Figure 1(a)shows the Brillouin scattering spectra of the ZPBx glasses.The Brillouin peak shifted to a higher frequency with anincrease in the substituent fraction. Figure 1(b) shows theBrillouin shift ¯B and Brillouin linewidth as a function ofB2O3 fraction. The molar volumes calculated from thedensities of these glasses decreased monotonically withincreasing x (Table 1). The decrease in molar volume indi-cates an increase in the packing density inside the glass.Because the elastic modulus reflects the packing density ofthe components and average bond strength,33,34) it is as-sumed that B2O3-substitution induces network connectiv-ity with denser packing. On the other hand, the linewidthof the ZPB10 glass is broader than that of other glasses.Since the linewidth reflects the scattering of ultrasonicwave in a matrix, it is assumed that B2O3-rich ZPB10 glasscontains wider structural distributions.As probes for amorphous materials, we used other ine-lastic light scattering spectroscopy and Raman scatteringspectra. The boson peak (BP) frequency, vBPmax, was ob-tained by fitting the Raman spectra in the BP with a log-normal function:35)»00ð¯Þ¯¼ Affiffiffiffiffiffi2³p·�exp � ðln ��®Þ22·2� �; ð1Þwhere ¯, e®, and · denote the frequency shift, median ofthe log-normal distribution, and standard deviation of ln v,respectively. The region of the Raman shift for fitting isfrom 0.6I0 at the lower frequency and 0.85I0 at the higherfrequency in the normalized spectra, where I0 denotes peakheight of the BPs. Figure 2(a) shows the Raman spectra ofthe ZPBx glasses (closed circles) and fitting curves (solidlines) obtained using a lognormal function. Figure 2(b)shows the BP energy vBPmax and peak height calculated inthe Stokes region of the Raman scattering of the ZPBxglasses as a function of the B2O3 fraction. With increasingB2O3 fraction, the vBPmax of the B2O3-substituted glasseslinearly increased. Contrarily, the peak height initially3231302928272610864201651601551501451403331292725Brillouin linewidth (MHz)Brillouin shift νB(GHz)B2O3 fraction (mol%)ShiftLinewidthFrequency (GHz)Intensity (arb. unit)ZP(x =0)ZPB5 ZPB10ZPB2.5ZPB1baFig. 1. Inelastic light scattering spectroscopy of B2O3substituted-ZP glasses I. a Brillouin scattering spectra of theZPBx glasses containing different B2O3 substitutions. b Brillouinshift ¯B (open circles) and the peak linewidth (closed squares) ofZPBx glasses as functions of the B2O3 fraction.Journal of the Ceramic Society of Japan 133 [1] 1-8 2025 JCS-Japan3decreases with increasing B2O3 fraction and then increasesin the ZPB10 glass. If the origin of the BP does notchange, it is expected to observe an inverse relationshipover the chemical composition range. This change in BPintensity suggests that the main origin of the BP haschanged with more B2O3 substitution.It is expected that the volume ratio and distribution ofcavities in glass correlates with the elastic properties.33) Inorder to quantify the cavity size, we measure positronannihilation measurement, in which the lifetime of ortho-positronium (o-Ps) reflects the infomration of cavities ininsulators.36) By fitting, three components can be obtained.The first component I1 is attributed to the lifetime of para-positronium (p-Ps), which has a theoretical value of 125ps. The second component I2 is the lifetime of positronannihilated without forming Ps, which also includes thepositron decay due to interaction with the Kapton film.The third component I3 is the reflected lifetime of o-Ps.Figure 3(a) shows the positron decay curves of the ZP andZPBx glasses. Decay profiles seem to changed dependingon the B2O3 fraction and the I3 ratio shown in Fig. 3(b) hasa negative correlation with the B2O3 fraction. Since I3ratios of SiO2 glass (standard),37) Li2O–SiO2 glass,37) andSrO–B2O3 glass,38) are approximately 50, 24–31, and 3–15%, the I3 ratio is considered to be an indicator for theelectron density of materials. However, the calculatedcavity radius [Fig. 3(b)] has a complicated change: theradius initially decreases with increasing B2O3 fraction andthen increases. It indicates that the structures detected byo-Ps have been changed.Considering the effect of B2O3 substitution on the struc-ture of ZP glasses, investigation of the coordination stateof Zn cations is important because it has been proposedthat the subnetwork structures of ZnOx units in ZnO–P2O5glasses affect their physical properties. In previous reports,the main ZP glass network comprised phosphate chainsand small ZnOx clusters. To examine the average coordi-nation state of Zn, we performed Zn K-edge XAFS mea-surements on ZP, ZPB5, and ZPB10 glasses. Figure 4(a)shows Zn K-edge XANES spectra of ZP, ZPB5, andZPB10 glasses along with that of ZnO. Although ZP andZPB5 glasses exhibit similar XANES spectra, the height ofwhite line decreases with increasing B2O3 fraction, whichcorrelates with the change of NWF from more ionic glassto more covalent glass. Figure 4(b) shows the k3»(k) ofthe Zn K-edge EXAFS spectra of these glasses. With anincrease in the substitution fraction, the amplitude of theoscillation decreased, indicating that the structural order-ing of the ZnOx units decreased with B2O3 substitution.This implies that the ZnOx units are distorted, or the com-plicated network structure suppresses the increase in oscil-lation amplitude. This indicates that the distorted ZnOnetwork was not sensitive to the type of substituted oxideunit. The Fourier transform of the EXAFS spectra of theZPBx glasses is shown in Fig. 4(c), along with that of thereference ZnO. The k-region for Fourier transform is fromν BPmax(cm−1)B2O3 fraction (mol%)Raman shift (cm−1)Intensity (arb. unit)ZP (x =0)baZPB5ZPB10ZPB2.5ZPB1Boson peak height (arb. unit)Boson peak heightνBPmax1501005005452504846444210864201.000.980.960.940.92Fig. 2. Inelastic light scattering spectroscopy of B2O3substituted-ZP glasses II. a Raman scattering spectra at theboson region of ZPBx glasses along with the nonsubstituted ZPglass. b Boson peak energy vBPmax and the peak height of ZPBxglasses as a function of the substitution fraction of B2O3.abTime (ns)Normalized intensityB2O3 fraction (mol%)I 3ratio (%)Cavity radius (nm)ZPZPB5ZPB2.5 ZPB1ZPB100.200.190.180.170.161086420864200.00010.0010.010.11141210864Fig. 3. Positron annihilation data of ZPBx glasses. a Posi-tron decay curves of ZPBx glasses along with ZP glass. b Cavityradius and the I3 fraction of ZPBx glass as a function of B2O3fraction.Masai et al.: Effect of borate substitution on zinc phosphate glassesJCS-Japan43.5 to 12¡¹1. The disordering caused by the substitutionof B2O3 causes an apparent decrease in the peak height ofthe first coordination shell. It is expected that structure ofNWF changes depending on the ZnOx units. To confirmthis hypothesis, we measured the local structure of theborate units using MAS NMR spectroscopy.We then examined the local coordination state of thesubstituted borate as a counterpart to the change in thephosphate units using MAS NMR spectra. Figure 5(a)shows the 11B MAS NMR spectra of ZPBx glasses. The11B MAS NMR spectra roughly comprise two units: four-coordinated boron BO4/2 located around 5 ppm and three-coordinated boron BO3/2 observed around 10 ppm.39–42)Although three-coordinated boron is further classified asring-structured three-coordinated boron BO3/2-ring andthree-coordinated non-ring boron BO3/2-non-ring, we treat1.51.00.50.09.689.679.669.651614121086420543210-4-2024121086420-15-10-5051015Wavenumber k (Å–1) k3χ(k) for glass (Å–3)|χ(R)| (Å–4)Radial distance (Å) k3χ(k) for ZnO (Å–3)ZPB5ZPZPB5ZPZnOZnObcaEnergy (keV) Normalized μ(E)ZPB10ZPB10ZPB5ZP ZnOZPB10Fig. 4. Substitution effect of B2O3 on the structure of Zncation. a Zn K-edge XANES spectra of 60ZnO–40P2O5 (ZP),ZPB5 and ZPB10 glasses along with that of ZnO. b Zn K-edgeEXAFS spectra of ZP, ZPB5 and ZPB10 glasses along with thatof ZnO. c Fourier-transform of EXAFS spectra of ZP, ZPB5, andZPB10 glasses along with that of ZnO.20 10 0 -10 -201009590858075108642010864203025201510501086420-60-40-20020B2O3 fraction (mol%)BO4/2 ratio (%)(ppm)Intensity (arb. unit) ZPB10ZPB5ZPB2.5cabdBO4/2amount (mol%)Intensity (arb. unit.)Chemical shift (ppm)Q2B2O3 fraction (mol%)Q0Q1Qnamount (mol%)ZPZPB5ZPB10ZPB2.5ZPB1Fig. 5. B2O3-substitution effect of glass structure inMAS NMR measurement. a 11B MAS NMR spectra of ZPBxcontaining different B2O3 fractions. b Change in the BO4/2 ratio(left) and BO4/2 amount (right) as a function of the B2O3 fraction.c 31P MAS NMR spectra of ZPBx glasses containing differentB2O3 fractions. The dashed lines indicate Q0, Q1, and Q2 units ofZPB10 glass after peak deconvolution. d Change in the Qnamount as a function of the B2O3 fraction.Journal of the Ceramic Society of Japan 133 [1] 1-8 2025 JCS-Japan5BO3/2 without further classification. If nearly all boroncations adopt the BO4/2 structure, it is expected that thenegative charge will be compensated for by the Zn cationnear the BO4/2 unit, as in ZPB2.5 glass. The presence ofBO3/2 is observed in the B2O3-rich ZPB10 glass. Thesespectra were deconvoluted using the dmfit 2002 program(dashed lines) to quantitatively discuss the structuralchanges in borate and to calculate each unit ratio.Figure 5(b) shows the BO4/2 ratio (left axis) and BO4/2amount (right axis) as functions of the B2O3 fraction. TheBO4/2 amount is obtained from the product of the BO4/2ratio and the B2O3 fraction. Although the BO4/2 amountincreases with increasing B2O3 fraction, the ratio of BO3/2units, which are loosely packed units compared with BO4/2,increases. Both the BO3/2 and BO4/2 units affect the glassnetwork to improve the thermal stability, as shown inTable 1. From 11B MAS NMR, we conclude that the B2O3species work in collaboration with the surrounding matrix,i.e. structural and electronic interactions with phosphatechains and ZnOx species.As previously reported, the main network structure ofZP glass consists of phosphate chains. Therefore, weexamined the structural changes in the phosphate unitsusing 31P MAS NMR spectroscopy. Figure 5(c) presentsthe 31P MAS NMR spectra of the ZPBx glasses. The31P MAS NMR spectra can be decomposed into threeunits: PO43¹ (Q0), PO3.52¹ (Q1), and PO3¹ (Q2).43) Thesuperscript n indicates the number of bridging oxygenatoms at a phosphorus cation. The dashed lines indicatethe units after peak deconvolution. The Q0 and Q1 unitsexhibited delocalized electrons caused by the conjugatedP–O¹ bonds. Unlike the Q2 chain, which forms the mainglass network, the Q0 and Q1 units cannot form a covalentmetal–oxygen chain structure.3–5) From the deconvolu-tion, we can discuss the structural change in the phosphateunit as a function of substitution. Figure 5(d) shows thenumber of Qn units, which are obtained by multiplying theratio of each Qn unit in each B2O3 fraction, as a function ofthe B2O3 fraction. The Q2 ratio decreases with increasingB2O3 fraction, whereas the Q0 and Q1 ratios increase. Thisindicates that the NWF borate units partially substitutedthe Q2 phosphate chain network. The Q2 amounts in boththe glasses also decreased.Here, we discuss the substitutional effect from the view-point of the number of bridging oxygen (BO) in the NWFspecies. Several units belong to the NWF groups: BO4/2¹,BO3/2 and PO3¹ (Q2 chain), and the numbers of BO ofeach unit are 4, 3, and 2, respectively. Based on the hy-pothesis that network formation is governed by such NWFunits, we calculate the total number of BO (NBO) of eachunit, which is obtained by the product of the amount ofeach unit and the ¯BO. Using the values in Figs. 5(b) and5(d), changes in NBO (¦NBO) of the B2O3-substituted sys-tem are plotted in Fig. 6. Because the absolute values ofthe increase (BO3/2 + BO4/2¹) are comparable to those ofthe decrease (Q2 chain) in the glass systems, the network-forming units of B2O3 work as a linkage of the networkinstead of the phosphate Q2 chains. Clear inverse relation-ships were observed between the increase in the NWF ofB2O3, suggesting that the decrease in the Q2 chain and theincrease in the NWF units of the B2O3 species are insepa-rable. Furthermore, the coordination states of the constit-uent B2O3 units change depending on the structure of thephosphate network. The obtained data also indicate thatthe structure of glass is to be discussed based on thecation-based network, not oxide-based composition.32)Although monolithic transparent ZPBx glasses wereobtained, we assume that a phase separation occurs, espe-cially in the ZPB10 glass. To confirm the tendency, wefocused on the X-ray scattering spectra. Figure 7(a) showsthe X-ray total structure factors S(Q) of the ZP and ZPBxglasses. The peak position of the first sharp diffractionpeak (FSDP)24,25) at ³1.8¡¹1 shifts slightly to the higherQ region by B2O3 substitution. Our group reported that anincrease in the ZnO fraction in ZP glass induces an in-crease in the packing density due to the formation of ZnOxpolyhedra to form glass network and the cleavage ofphosphate chains to be isolated phosphate structures.22)This shift indicates that the periodicity of the networkstructure, which is the origin of the FSDP, becomes nar-rower, that is, a densely packed structure is generated bysubstitution. It is notable that the S(Q) increases slightlyat <1¡¹1, especially in ZPB10 glass. Figure 7(b) showssmall-angle X-ray scattering spectra of ZPBx glasses alongwith that of ZP glass. An increase in the intensity in thelow-Q region, which is an indicator of phase separation, isclearly observed for the ZPB10 glasses. Similar to theresults of X-ray diffraction, the ZPB10 glass exhibits thelargest small-angle components, which is consistent with aprevious phase separation behavior in Sn2+-doped ZnO–P2O5–B2O3 glass.18) It suggests that small-angle scatteringmeasurement is useful for understanding of phase separa-tion of glass,44) including precipitation of nanocrystallites.Based on the results, we confirm that atomistic phaseseparation occurs in ZnO–P2O5–B2O3 glass with highB2O3 substitution. Although apparent macroscopic phaseseparation was not observed in the B2O3-rich glass, atom-istic aggregation is detected by several analytical methods.Since atomistic phase separation is thought be an initialNBO (arb. unit)NBO(PO3−)NBO(BO4/2 + BO3/2)B2O3 fraction (mol%)-40-20020401086420Fig. 6. Change in the number of borate units and PO4/2 (Q2)unit as a function of the B2O3 fraction. The absolute values ofthe increase ¦NBO are comparable to those of decrease ¦NBO.Masai et al.: Effect of borate substitution on zinc phosphate glassesJCS-Japan6stage of macroscopic phase separation, the approach willhelp to understand generation of heterogeneous regions inglass, including precipitation of crystallites.4. SummaryWe examined the structure and physical properties ofB2O3-substituted 60ZnO–40P2O5 glasses. The substitutedB2O3 species work as network-forming units that affectclosed packing and increase the elastic properties. By com-paring the network-forming bridging oxygen, it is sug-gested that the change in the phosphate chains is a coun-terpart to the change in the borate unit to maintain theglass-forming network. Although phase separation is notnotable at lower B2O3-substitution, the formation of clus-ters is observed in B2O3-rich ZP glasses. The formation ofcluster observable in small-angle diffraction is consistentwith other measurement data in inelastic light scatteringand 11B MAS NMR. The formation of heterogeneousregion is thought to be atomistic phase separation, whichwill be a key for understanding of structural change inglass.Acknowledgements This work was partially supportedby the Japan Society for the Promotion of Science Grant-in-Aid for Scientific Research (B) Numbers 18H01714,22H01785 (H.M.), 21H01018 (Y.F. & A.K.), for ScientificResearch (S) Number 19H05618 (A.K.), for ScientificResearch (C) Number 24K08045 (Y.F.), and for Transforma-tive Research Areas (A) Numbers 20H05881 (Y.O.),20H05882 (H.M.), and 23H04096 (T.O.). This work wasalso supported by the Tokyo Ohka Foundation for ThePromotion of Science and Technology. High-energy XRDmeasurements were performed on the BL04B2 beamline atSPring-8 with the approval of the Japan Synchrotron Radia-tion Research Institute (JASRI) (Proposal Nos. 2017B0134and 2024A1023). Zn K-edge XAFS measurements wereperformed on the BL01B1 beamline at SPring-8 with theapproval of the JASRI (Proposal No. 2021A1221). We appre-ciate the experimental support in XAFS and high-energyXRD measurements by Drs. T. Ina (JASRI) and S. Kohara(NIMS).Data Availability The authors declare that all relevantdata supporting the findings of this study are available fromthe corresponding author upon request.References1) W. H. Zachariasen, J. Am. Chem. Soc. 54, 3841 (1932).2) K. H. Sun, J. Am. Ceram. Soc. 30, 277 (1947).3) J. Brow, J. Non-Cryst. Solids 263–264, 1 (2000).4) U. Hoppe, J. Non-Cryst. Solids 195, 138 (1996).5) R. K. Brow, J. Non-Cryst. Solids 194, 267 (1996).6) R. K. Brow, D. R. Tallant, S. T. Myers and C. C. Phifer,J. Non-Cryst. Solids 191, 45 (1995).7) S. W. Martin, Eur. J. Sol. State Inor. 28, 163 (1991).8) G. Walter, U. Hoppe, J. Vogel, G. Carl and P. Hartmann,J. Non-Cryst. Solids 333, 252 (2004).9) J. W. Wiench, M. Pruski, B. Tischendorf, J. U. Otaigbeand B. C. Sales, J. Non-Cryst. Solids 263–264, 101(2000).10) U. Hoppe, R. Kranold, A. Barz, D. Stachel, J. Neuefeindand D. A. Keen, J. Non-Cryst. Solids 293–295, 158(2001).11) G. Walter, U. Hoppe, T. Baade, R. Kranold and D.Stachel, J. Non-Cryst. Solids 217, 299 (1997).12) K. Suzuya, K. Itoh, A. Kajinami and C.-K. Loong,J. Non-Cryst. Solids 345–346, 80 (2004).13) P. Sengupta, J. Hazard. Mater. 235–236, 17 (2012).14) P. I. Paulose, G. Jose, V. Thomas, N. V. Unnikrishnanand M. K. R. Warrier, J. Phys. Chem. Solids 64, 841(2003).15) J. H. Campbell and T. I. Suratwala, J. Non-Cryst. Solids263–264, 318 (2000).16) H. Masai, J. Ceram. Soc. Jpn. 121, 150 (2013).17) D. Pugliese, N. G. Boetti, J. Lousteau, E. Ceci-Ginistrelli, E. Bertone, F. Geobaldo and D. Milanese,J. Alloy. Compd. 657, 678 (2016).18) H. Masai, S. Okumura, T. Ohkubo and T. Yanagida,Opt. Mater. Express 7, 2993 (2017).19) M. Shiga, A. Hirata, Y. Onodera and H. Masai,Commun. Mater. 4, 91 (2023).20) S. K. Lee, H. N. Kim, B. H. Lee, H.-I. Kim and E. J.Kim, J. Phys. Chem. B 114, 412 (2010).21) G. Bonny, D. Terentyev, L. Malerba and D. Van Neck,Phys. Rev. B 79, 104207 (2009).22) Y. Onodera, S. Kohara, H. Masai, A. Koreeda, S.Okamura and T. Ohkubo, Nat. Commun. 8, 15449(2017).23) U. Hoppe, G. Walter, G. Carl, J. Neuefeind and A. C.Hannon, J. Non-Cryst. Solids 351, 1020 (2005).403020100.50.40.30.20.11.41.21.00.80.60.40.20.0-0.22.52.01.51.00.50.0Q (Å–1)S(Q)Q (Å–1)Intensity (arb. unit)abZPZPB5ZPB10ZPB2.5ZPB1ZPZPB5ZPB2.5ZPB1ZPB100.04Fig. 7. Structural analysis using synchrotron X-rays. aStructure factor S(Q) of ZPBx glasses along with that of ZP glass.b Small-angle X-ray scattering spectra of ZPBx glasses alongwith that of ZP glass.Journal of the Ceramic Society of Japan 133 [1] 1-8 2025 JCS-Japan724) Y. Onodera, J. Ceram. Soc. Jpn. 130, 627 (2022).25) S. Kohara, J. Ceram. Soc. Jpn. 130, 531 (2022).26) H. Masai, J. Non-Cryst. Solids: X 15, 100105 (2022).27) H. Masai, T. Tanimoto, T. Fujiwara, S. Matsumoto, Y.Takahashi, Y. Tokuda and T. Yoko, J. Non-Cryst. Solids358, 265 (2012).28) A. Koreeda and S. Saikan, Rev. Sci. Instrum. 82, 126103(2011).29) M. Yamawaki, Y. Kobayashi, K. Hattori and Y.Watanabe, Jpn. J. Appl. Phys. 50, 086301 (2011).30) S. Kohara, M. Itou, K. Suzuya, Y. Inamura, Y. Sakurai,Y. Ohishi and M. Takata, J. Phys.-Condens. Mat. 19,506101 (2007).31) B. Reval and M. Newville, J. Synchrotron Radiat. 12,537 (2005).32) H. Masai, T. Mihara and K. Kintaka, J. Ceram. Soc.Jpn. 131, 466 (2023).33) T. Rouxel, J. Ceram. Soc. Jpn. 130, 519 (2022).34) A. Masuno, J. Ceram. Soc. Jpn. 130, 563 (2022).35) V. K. Malinovsky, N. Novikov and A. P. Sokolov, Phys.Lett. A 153, 63 (1991).36) R. A. Pethrick, Prog. Polym. Sci. 22, 1 (1997).37) K. Inoue, H. Kataoka, Y. Nagai, M. Hasegawa and Y.Kobayashi, J. Appl. Phys. 115, 204903 (2014).38) H. Masai, T. Ohkubo, Y. Fujii, A. Koreeda, T. Yanagida,T. Ina and K. Kintaka, Sci. Rep.-UK 11, 3811 (2021).39) J. F. Stebbins, P. Zhao and S. Kroeker, Solid State Nucl.Mag. 16, 9 (2000).40) P. Zhao, S. Kroeker and J. F. Stebbins, J. Non-Cryst.Solids 276, 122 (2000).41) M. Vasilescu and S. Simon, Mod. Phys. Lett. B 16, 423(2002).42) S. Sen, S. Xu and J. F. Stebbins, J. Non-Cryst. Solids226, 29 (1998).43) R. J. Kirkpatrick and R. K. Brow, Solid State Nucl. Mag.5, 9 (1995).44) G. N. Greaves, M. C. Wilding, S. Fearn, D. Langstaff, F.Kargl, S. Cox, Q. Vu Van, O. Majérus, C. J. Benmore,R. Weber, C. M. Martin and L. Hennet, Science 322,566 (2008).Masai et al.: Effect of borate substitution on zinc phosphate glassesJCS-Japan8