Skip to Main Content
Skip Nav Destination

In recent years, as new fields in biomedical applications have emerged, considerable attention has been devoted to glass materials for their ability to promote bone formation and their application for the treatment and repair of soft tissue. So far, silicate-based glasses have been widely applied as biomedical glasses. In some cases, however, silicate-based glasses are not suitable for all applications. Phosphate-based and borate-based glasses are proposed as attractive alternatives. They are prospective biomaterials, because their structures are controllable and their characteristic ion-releasing behaviour can be tuned.

Phosphate glasses have structural features that are significantly different from those of silicate glasses. Phosphate glasses typically consist of a layered or chain structure in which PO4 units are connected and, with a few exceptions, do not form a three-dimensional PO4 network. This is because P is pentavalent and can form a 4-coordinated structure with oxygen by forming a dπ–pπ-type bond with one of the four coordinating oxygens. Therefore, non-bridging oxygens (NBOs) are naturally present in phosphate glasses, eliminating the requirement of introducing modifier components such as alkali oxides. The introduction of modifiers leads to the formation of various phosphate groups such as orthophosphate units, Q0, pyrophosphate units, Q1, metaphosphate (intermediate) units, Q2, and branching units, Q3. The difference in π-bonding and σ-bonding properties of each group influences the chemical properties of the phosphate glasses. Q3 is chemically unstable because the double bond moiety is localised to a single P–O bond, and it easily reacts with water to convert to Q2 delocalised to two P–O bonds. This is called the “anti-branching rule”.1  This delocalisation of the double bond can be verified through X-ray photoelectron spectroscopy (XPS) by using the O 1s spectra2  and via electron spin resonance (ESR) spectroscopy.3  In M2O–P2O5 (M = alkali metal cation) glasses, upon increasing the M2O content to a composition at which Q3 groups disappear, such glasses become more chemically stable, in contrast to conventional silicate glasses. When glasses with an ultraphosphate composition are melted, moisture in the raw materials and in the atmosphere is taken up. The (M2O + H2O)/P2O5 ratio tends to approach unity, following the “anti-branching rule”, to eliminate Q3 branching groups. As a result, phosphate glasses include a larger number of OH groups than silicate or borate glasses do.4 

Because phosphate glasses have lower electron densities on oxygen than silicate glasses and are more acidic, they can be used as good solvents for basic oxides and serve as hosts with weaker ligand fields. Moreover, it is considered that these structural features also contribute to the good glassification ability, even when the P2O5 amount is 40% or less, and to the dissolution of oxide compounds, such as Ag2O, at high concentrations. The chemical durability of phosphate glasses can be considerably improved by incorporating B2O3 and/or Al2O3 to improve the symmetry around tetrahedral PO4, and various practical glasses such as ultraviolet transparent glasses or sulfur-resistant glasses have been developed. The detailed structure and properties of phosphate glasses for biological use are described by Delia Brauer in Chapter 2.

The structure of phosphate glasses has been described in many reviews,5–8  including Van Wazer9  and Abe10 et al. Therefore, in this chapter, the author will introduce recently reported phosphate glasses with interesting specific structures and properties.

Unlike silicate and borate glasses, phosphate glasses can be vitrified even when their composition comprises a small amount of so-called “network former” (NWF) ions. When appropriate compositions are chosen, phosphate glasses consisting of Q0 and/or Q1 units can be obtained without polymerisation of PO4 tetrahedra.8,10  Glass “networks” are formed by PO4 tetrahedra being ionically connected via metal cations. In other words, it is difficult to form a glassy state only with the Q0 and/or Q1 units; however, phosphate glasses can be vitrified by incorporating other components that play the role of a network former. These glasses are examples of systems in which Zachariasen's “random network model” cannot be applied, and are classified as “invert glasses”, as proposed by Trap and Stevels.11  Here, independent anionic groups are linked by cations that were originally network modifier (NWM) components; the anionic groups act as NWM-like groups and the cations are functionally considered to be NWF. For biomedical application, 60CaO–30P2O5–10TiO2 (in mol%) glass12  has been reported as one of the ideal glasses. This glass exhibits in vitro bioactivity (apatite-forming ability in a simulated body fluid). This glass and its glass-ceramic with a modified composition have been reported to adhere directly to bone in animal experiments.13 

Maeda et al.14  have clarified the effect of TiO2 in a calcium phosphate invert glass on the local structure by using spectroscopic methods and simulations based on molecular dynamics (MD). From 31P magic angle spinning nuclear magnetic resonance (MAS-NMR) spectra, Q1 and Q2 peaks were observed in the 60CaO–40P2O5 glass, whereas Q0 and Q1 peaks were observed in the 60CaO–30P2O5–10TiO2 glass. Malavasi et al.15  reported the network connectivities for 60CaO–40P2O5 and 60CaO–30P2O5–10TiO2 glasses to be 1.42 and 0.77, respectively. The shortening of the phosphate chain structure by incorporating TiO2 was considered to cause a decrease in network connectivity. In the Raman spectrum of the 60CaO–30P2O5–10TiO2 glass, peaks attributed to the TiO4 and TiO6 units were observed.

In the XPS O1s spectrum of the 60CaO–40P2O5 glass, peaks attributed to P–O–P and P–O–Ca bonds were observed, while in that of the 60CaO–30P2O5–10TiO2 glass, the P–O–Ti peak was observed in addition to the P–O–P and P–O–Ca peaks. These results suggest that in the 60CaO–30P2O5–10TiO2 glass, 4- and 6-fold coordinated titanium ions exist and form P–O–Ti bonds with the phosphate species.

Figure 1.1 shows the structural model of the glasses obtained using the MD method. In this figure, the structures wherein P, Ti and O are contributing to the glass network structure are extracted. The network connectivities for 60CaO–40P2O5 and 60CaO–30P2O5–10TiO2 glasses were 1.53 and 0.67, respectively, which are close to the spectroscopy results. Because the formation of the P–O–Ti bond is considered in the XPS measurement, TiO2 also acts as an NWF. Assuming that P2O5 and TiO2 work to form the network structure of the 60CaO–30P2O5–10TiO2 glass, its connectivity is 1.81, which is higher than that for the 60CaO–40P2O5 glass.

Figure 1.1

Snapshots of (a) 60CaO–40P2O5 and (b) 60CaO–30P2O5–10TiO2 glass models simulated using 2000 atoms. Ca atoms are not shown for easy viewing. Colour legend: P (purple), Ti (green), and O (red bar).

Figure 1.1

Snapshots of (a) 60CaO–40P2O5 and (b) 60CaO–30P2O5–10TiO2 glass models simulated using 2000 atoms. Ca atoms are not shown for easy viewing. Colour legend: P (purple), Ti (green), and O (red bar).

Close modal

Water wettability, which is a functional indicator for biomaterials, of 60CaO–40P2O5 and 60CaO–30P2O5–10TiO2 glasses was evaluated to be 75.0° ± 4.4° and 54.0° ± 4.9°, respectively. The incorporation of TiO2 increases the network connectivity of the glass, resulting in an increase in the bond strength of the entire glass. This increases the surface tension, which contributes to the improvement of hydrophilicity. The elucidation of the local structure of phosphate glasses has allowed its correlation with their biological functions.

The development of this invert glass has led to the investigation of new biomedical glasses, which are described in detail by Sungho Lee in Chapter 5.

It has been reported that the incorporation of TiO2 and CaO improves the chemical durability of phosphate glasses, thus widening their scope for biomedical application.16,17  Recently, silicate ions have been recognised as therapeutic ions that effectively promote bone formation.18,19  In some phosphate glasses, chemical durability may be improved by incorporating SiO2 into the composition. If silicate and/or Ca2+ ions can be released slowly, silicophosphate glasses are expected to be useful as biomedical glasses.

The oxygen coordination number of silicon in oxide glasses is usually 4; that is, a tetrahedral SiO4 structure is formed. However, Dupree et al.20  revealed that in phosphate glasses with a certain composition (P2O5 > 40 mol%), part of silicon forms an octahedral 6-fold coordinated structure even under atmospheric pressure. Such structure formation was promoted by the decrease in the electron-donating property (optical basicity) of O atoms in the glasses due to the presence of a large amount of P2O5 with high acidity.21  When a monovalent element such as lithium or sodium and/or a divalent element such as magnesium or calcium are incorporated in some silicophosphate glasses, the negative charge of the 6-fold coordinated silicate structure, SiO6, is compensated. Therefore, this structure tends to form.22 

The formation of the SiO6 structure in silicophosphate glasses has been reported to affect their thermomechanical properties23  and optical properties24  and improve their durability in water;25  the structure may be effective in controlling various properties. The improvement of properties is attributed to the bridging of the phosphate chains in the glasses by the SiO6 structure. In contrast, the SiO4 structure in the glasses incorporates the phosphate chain structure and breaks the chain, for example, resulting in reduced chemical stability. Consequently, the durability of the glasses in water deteriorates.

A structural analysis of the phosphate glasses containing SiO6 units has been performed by MAS-NMR26,27  and/or X-ray absorption fine structure (XAFS)28  methods. Theoretical calculations of SiO6 structures in phosphate glasses using a density functional theory (DFT) method have revealed the presence of atomic clusters around silicon at a scale of approximately a dozen atoms.22  Recently, the author of this chapter reported a glass model analysed by MD simulation.29 

Figure 1.2 shows an example of the calculation results for the P2O5–SiO2–Na2O glass. It was found that the SiO6 structure contributes to glass network formation and converts the QP2 unit into an energetically stable QP3 unit; this is favoured because the Si–O bond in the SiO6 structure has higher ionicity than the Si–O bond in SiO4 and the P–O bonds. It was also confirmed that the QP3 unit is likely to be preferentially coordinated to the SiO6 structure, and Na+ ions with high coordination numbers are easily coordinated in the vicinity by the interaction with O of the Si–O–P bond. The formation of a SiO6 structure may lead to better control of the hydrolysis reaction of QP3 units and the diffusion of Na+ ions. The effective incorporation of SiO6 and SiO4 structures in phosphate glasses is likely to be a critical step in the development of new biomaterials.

Figure 1.2

A snapshot of a 55.0P2O5–21.3SiO2–23.7Na2O glass model simulated using 510 atoms. Colour legend: P (purple), Si (ivory), Na (yellow), and O (red).

Figure 1.2

A snapshot of a 55.0P2O5–21.3SiO2–23.7Na2O glass model simulated using 510 atoms. Colour legend: P (purple), Si (ivory), Na (yellow), and O (red).

Close modal

Borate glasses are known to have various groups of medium-range structures (borate superstructural units). As a particularly specific and well-known property, the so-called “boron oxide anomaly” phenomenon has been extensively investigated.30,31  For example, by increasing the content of an alkali oxide (M2O; M = Li, Na, K, Rb, Cs) in M2O–B2O3 glasses, their thermal expansion coefficients decrease and reach a minimum value at 16 mol%. Krogh-Moe32  considered that a B2O3 glass consists only of a 3-membered ring (boroxol ring) of 3-fold coordinated boron and proposed a model in which the boroxol ring changes to a ring with a 4- or more membered ring when an alkali oxide is incorporated. The boron oxide anomaly of borate glasses consisting of these rings can be clearly explained. In an MD study before a report by Yamamoto et al.,33  the number of the boroxol rings was less than or almost zero compared with that determined by the experimental results.34,35  However, they clarified the high reproducibility of the boron oxide anomaly and the relationship between the ring structures and the change in their thermal expansion coefficients by performing an MD simulation on the basis of the interatomic interaction estimated from the first-principles calculation.

One of the attractive properties of borate glasses as a biomaterial is their variation of chemical durability. Borosilicate glasses are chemically durable, while typical borate glasses exhibit rapid dissolution in aqueous solutions.

Many aspects of the structure of borate glasses remain unknown, and to the best of our knowledge, such glasses have few practical applications. Nevertheless, borate, borosilicate, and borophosphate glasses have been developed and evaluated to obtain properties that are difficult to acquire with silicate and phosphate glasses. Boron oxide plays an important role in adjusting the solubility of ion-releasing-type multi-component glasses.

Boron has been reported to influence the activity of osteoblasts and promote angiogenesis.36,37  Boron deficiency sometimes adversely affects bone formation;38  boron deficiency in rats has been reported to reduce bone volume and thickness and to impair the regeneration of the alveolar bone.39  Therefore, some studies have been conducted to investigate the application scope of boron-based glasses in biological functions. Unary borate glass exhibits extremely low chemical durability. However, its durability improves drastically by incorporating SrO in its composition, for example.40  This is because the ionic radius of Sr2+ is larger than that of Ca2+. Elements with large ionic radii require more space in the glass network, thus controlling the mobility of other ions. The Sr2+ ion is known as a therapeutic ion that contributes to bone formation.

For widening the application scope of borate-based biomaterials, suppression of their solubility is essential. However, their high solubility can promote some applications. The use of borate glass corresponding to Bioglass® 45S5 (46.1B2O3–26.9CaO–24.4Na2O–2.6P2O5 (mol%); wherein SiO2 is completely replaced by B2O3) leads to the formation of a larger amount of hydroxyapatite (HA) in simulated body fluid (in vitro bioactivity) than that when original silicate-based Bioglass® 45S5 is used.41  When the borate glass was soaked in a simulated body fluid, Ca2+, PO43−, Na+, and BO33− ions from the glass were released, and PO43− ions in the solution reacted with Ca2+ ions to form HA on the glass surface. The starting point that induces HA nucleation is an attack on the B–O network structure in an aqueous solution. It is considered that this is because the BO3 structure cannot completely form a strong three-dimensional network, unlike the SiO4 structure, thus inducing the continuous dissolution of the glass. In animal experiments in which 45S5-based glass particles containing B2O3 were implanted in the tibia and contralateral tibia of male Wistar rats, a statistically significant increase in bone tissue formation around the particles was observed.42 

Borate glasses, as well as phosphate-based glasses, are excellent solvents/matrices and can be used for the release of therapeutic ions.19  The use of borate glasses containing several therapeutic ions to heal bone defects has been investigated for promoting bone regeneration. For example, when borate glasses containing divalent cations such as Mg2+, Ca2+, and Sr2+ ions were used, vigorous cell adhesion and growth were observed along with high solubility in water.43  Borate glasses may stimulate bone marrow mesenchymal stem cells to form bone.

It has also been reported that boron is involved in promoting wound healing44  and releasing growth factors and cytokines.45  The dissolution of a large amount of Ca2+ ions from borate glasses supported the migration of epidermal cells to wounds and contributed to the regeneration of skin and soft tissues.46  Therefore, one of the most important current applications of borate glasses is as a wound healing system.

Phosphate-based and borate-based glasses are interesting biomaterials that exhibit unique behaviour because of their structures and physical properties. While silicate glasses tend to have a three-dimensional network structure composed of SiO4 units, phosphate glasses often have chain structures composed of PO4 units. Recently, a “mixed alkaline metaphosphate glass”47  showing entropic elasticity, which mimics the organic chain polymer structure, has been reported. When this glass is stretched unidirectionally while being heated and then cooled, the –P–O–P– chains are linearly arranged along the stretching direction; therefore, the glass shows a large anisotropy. When reheating this glass at the stretching temperature, the glass shrinks. Even oxide glasses, which are hard and fragile at room temperature, can exhibit the property of expanding and contracting like rubber at a high temperature if the internal structure at atomic and molecular scales is controlled well. As described above, phosphate glasses have intermediate properties between inorganic and organic matter, and there is still significant potential for discovering new phenomena and functions.

One of the unique properties of borate glasses is their high solubility in aqueous solutions. Many studies have been conducted to utilise this property to explore their potential for promoting cell adhesion, proliferation, and differentiation. Modifying biomaterials to allow functions such as angiogenesis and soft tissue infiltration and exploring new mechanisms for promoting bone formation are challenging and require further investigation.

The structural analysis of silicate-based glasses has been widely conducted for promoting their application as biomedical glasses. While groups are conducting the necessary analyses of phosphate-based and borate-based glasses, further advances are expected in the future. Slow progress on finding new functions by using computational-chemistry-based approaches, such as simulation of the interface between the living tissue and material, has been made. Finally, digital transformation of biomedical glasses is a critical future research topic.

1.
van Wazer
 
J. R. V.
Holst
 
K. A.
J. Am. Chem. Soc.
1950
, vol. 
72
 pg. 
639
 
2.
Gresch
 
R.
Müller-Warmuth
 
W.
Dutz
 
H.
J. Non-Cryst. Solids
1979
, vol. 
34
 pg. 
127
 
3.
Berger
 
R.
Vignaud
 
G.
Olazcuaga
 
R.
Zahir
 
M.
J. Non-Cryst. Solids
1983
, vol. 
54
 pg. 
113
 
4.
Abe
 
Y.
Hosono
 
H.
Ohta
 
Y.
Hench
 
L. L.
Phys. Rev. B
1988
, vol. 
38
 pg. 
10166
 
5.
Martin
 
S. W.
Eur. J. Solid State Inorg. Chem.
1991
, vol. 
28
 pg. 
163
 
6.
Brow
 
R. K.
J. Non-Cryst. Solids
2000
, vol. 
263–264
 pg. 
1
 
7.
Hoppe
 
U.
Walter
 
G.
Kranold
 
R.
Stachel
 
D.
J. Non-Cryst. Solids
2000
, vol. 
263–264
 pg. 
29
 
8.
J. R.
van Wazer
,
Phosphorus and its Compounds
,
Interface Science
,
New York
,
1958
,
vol. 1
, p. 419
9.
Y.
Abe
,
Topics in Phosphorus Chemistry
, ed. M. Grayson and E. J. Griffith,
Wiley
,
New York
,
1983
,
vol. 11
, pp. 19–67
10.
A. E. R.
Westman
,
Modern Aspects of the Vitreous State
, ed. J. D. Mackenzie,
Butterworth
,
1960
,
vol. 1
, pp. 63–91
11.
Trap
 
H. J. L.
Stevels
 
J. M.
Phys. Chem. Glasses
1960
, vol. 
1
 pg. 
107
 
12.
Kasuga
 
T.
Acta Biomater.
2005
, vol. 
1
 pg. 
55
 
13.
Kasuga
 
T.
Hattori
 
T.
Niinomi
 
M.
Phosphorus Res. Bull.
2012
, vol. 
26
 pg. 
8
 
14.
Maeda
 
H.
Tamura
 
T.
Kasuga
 
T.
J. Phys. Chem. B
2017
, vol. 
121
 pg. 
5433
 
15.
Malavasi
 
G.
Pedone
 
A.
Menziani
 
M. C.
J. Phys. Chem. B
2013
, vol. 
117
 pg. 
4142
 
16.
Uo
 
M.
Mizuno
 
M.
Kuboki
 
Y.
Makishima
 
A.
Watari
 
F.
Biomaterials
1998
, vol. 
19
 pg. 
2277
 
17.
Kiani
 
A.
Lakhkar
 
N. J.
Salih
 
V.
Smith
 
M. E.
Hanna
 
J. V.
Newport
 
R. J.
Pickup
 
D. M.
Knowles
 
J. C.
Philos. Trans. R. Soc. London, Ser. A
2012
, vol. 
370
 pg. 
1352
 
18.
Xynos
 
I. D.
Edgar
 
A. J.
Buttery
 
L. D.
Hench
 
L. L.
Polak
 
J. M.
Biochem. Biophys. Res. Commun.
2000
, vol. 
276
 pg. 
461
 
19.
Hoppe
 
A.
Güldal
 
N. S.
Boccaccini
 
A. R.
Biomaterials
2011
, vol. 
32
 pg. 
2757
 
20.
Dupree
 
R.
Holland
 
D.
Mortuza
 
M. G.
Nature
1987
, vol. 
328
 pg. 
416
 
21.
Yamashita
 
H.
Yoshino
 
H.
Nagata
 
K.
Yamaguchi
 
I.
Ookawa
 
M.
Maekawa
 
T.
J. Ceram. Soc. Jpn.
1998
, vol. 
106
 pg. 
539
 
22.
Miyabe
 
D.
Takahashi
 
M.
Tokuda
 
Y.
Yoko
 
T.
Uchino
 
T.
Phys. Rev. B
2005
, vol. 
71
 pg. 
172202
 
23.
Zeng
 
H.
Jiang
 
Q.
Liu
 
Z.
Li
 
X.
Ren
 
J.
Chen
 
G.
Liu
 
F.
Peng
 
S.
J. Phys. Chem. B
2014
, vol. 
118
 pg. 
5177
 
24.
Zeng
 
H.
Jiang
 
Q.
Li
 
X.
Feng
 
F.
Tian
 
T.
Zhang
 
H.
Chen
 
G.
Appl. Phys. Lett.
2015
, vol. 
106
 pg. 
021903
 
25.
Miura
 
A.
Obata
 
A.
Kasuga
 
T.
Phosphorus Res. Bull.
2022
, vol. 
38
 pg. 
1
 
26.
Ren
 
J.
Eckert
 
H.
J. Phys. Chem. C
2018
, vol. 
122
 pg. 
27620
 
27.
Knight
 
C. T. G.
Kirkpatrick
 
R. J.
Oldfield
 
E.
J. Non-Cryst. Solids
1990
, vol. 
116
 pg. 
140
 
28.
Ide
 
J.
Ozutsumi
 
K.
Kageyama
 
H.
Handa
 
K.
Umesaki
 
N.
J. Non-Cryst. Solids
1966
, vol. 
353
 pg. 
2007
 
29.
K.
Takada
,
T.
Tamura
,
H.
Maeda
,
A.
Obata
and
T.
Kasuga
,
Preprint of Annual Meeting of Ceramic Society of Japan
,
Tokyo
,
vol. 2D04
,
2020
30.
Zhong
 
J.
Bray
 
P. J.
J. Non-Cryst. Solids
1989
, vol. 
111
 pg. 
67
 
31.
Shelby
 
J. E.
J. Am. Ceram. Soc.
1983
, vol. 
66
 pg. 
225
 
32.
Krogh-Moe
 
J.
J. Non-Cryst. Solids
1969
, vol. 
1
 pg. 
269
 
33.
Yamamoto
 
Y.
Sawaguchi
 
N.
Sasaki
 
M.
J. Comput. Chem., Jpn.
2018
, vol. 
17
 pg. 
113
  
, in Japanese.
34.
Xu
 
Q.
Kawamura
 
K.
Yokokawa
 
T.
J. Non-Cryst. Solids
1988
, vol. 
104
 pg. 
261
 
35.
Takada
 
A.
Catlow
 
C. R. A.
Price
 
G. D.
J. Phys.: Condens. Matter
1995
, vol. 
7
 pg. 
8659
 
36.
Pizzorno
 
L.
Integr. Med. Clin. J.
2015
, vol. 
14
 pg. 
35
 
37.
Nielsen
 
F. H.
Nutrition
2000
, vol. 
16
 pg. 
512
 
38.
Balasubramanian
 
P.
Büttner
 
T.
Pacheco
 
V. M. M.
Boccaccini
 
A. R.
J. Eur. Ceram. Soc.
2018
, vol. 
38
 pg. 
855
 
39.
Nielsen
 
F. H.
Stoecker
 
B. J.
J. Trace Elem. Med. Biol.
2009
, vol. 
23
 pg. 
195
 
40.
Pan
 
H. B.
Zhao
 
X. L.
Zhang
 
X.
Zhang
 
K. B.
Li
 
L. C.
Li
 
Z. Y.
Lam
 
W. M.
Lu
 
W. W.
Wang
 
D. P.
Huang
 
W. H.
Lin
 
K. L.
Chang
 
J.
J. R. Soc., Interface
2010
, vol. 
7
 pg. 
1025
 
41.
Huang
 
W.
Day
 
D. E.
Kittiratanapiboon
 
K.
Rahaman
 
M. N.
J. Mater. Sci.: Mater. Med.
2006
, vol. 
17
 pg. 
583
 
42.
Gorustovich
 
A.
Guglielmotti
 
M. B.
López
 
J. M. P.
Cabrini
 
R. L.
Key Eng. Mater.
2005
, vol. 
284–286
 pg. 
913
 
43.
Fernandes
 
J. S.
Gentile
 
P.
Crawford
 
A.
Pires
 
R. A.
Hatton
 
P. V.
Reis
 
R. L.
Tissue Eng., Part A
2017
, vol. 
23
 pg. 
1331
 
44.
Yang
 
Q.
Chen
 
S.
Shi
 
H.
Xiao
 
H.
Ma
 
Y.
Mater. Sci. Eng. Carbon
2015
, vol. 
55
 pg. 
105
 
45.
Dzondo-Gadet
 
M.
Mayap-Nzietchueng
 
R.
Hess
 
K.
Nabet
 
P.
Belleville
 
F.
Dousset
 
B.
Biol. Trace Elem. Res.
2002
, vol. 
85
 pg. 
23
 
46.
Liu
 
X.
Rahaman
 
M. N.
Day
 
D. E.
J. Mater. Sci.: Mater. Med.
2013
, vol. 
24
 pg. 
583
 
47.
Inaba
 
S.
Hosono
 
H.
Ito
 
S.
Nat. Mater.
2015
, vol. 
14
 pg. 
312
 
Close Modal

or Create an Account

Close Modal
Close Modal