Sn-containing silicate glasses with zero photoelastic constant (PEC) can potentially substitute zero-PEC Pb-containing glasses as optical fiber current sensor components based on the Faraday effect. These compounds allow monitoring of the electric power by measuring the electric current in high-voltage conductors operated with a 1550-nm light. The toxicity of Pb in these glasses still remains an important issue. However, replacing Pb in the sensors while minimizing the PEC of the resulting device would represent a significant breakthrough. We report a 43.5SnO–56.5SiO2 glass in molar% with zero PEC of + 0.01 × 10−12 Pa−1 observed with a wavelength of 632.8 nm.
© 2017 Optical Society of America
The optical properties of glasses having zero photoelastic constant (PEC) are particularly useful for maintaining the polarization of light under external mechanical stress and heat conditions in polarization or interference optical devices such as beam splitters, polarizers, and lenses . Additional optical devices fabricated with zero-PEC glasses are required for optical-fiber-type current sensors based on the Faraday effect. These devices are to be used in systems monitoring the current in electrical conductors. These systems may include the basic technology for the control, protection, and monitoring of electric power generation plants, factories, and railways . Despite showing very low PEC values , fiber-type current sensors made of PbO–SiO2 glasses are not sustainable because of the toxicity of Pb.
Thus, environmentally friendly fiber-type current sensors are desirable for sustainable applications involved in the upcoming smart grid society. One purpose of this society is the suppression of the total energy including electric power. New Pb-free fiber-type optical current sensors with appropriate glass compositions are required. We studied Pb-free glass compositions (i.e., ZnO–SnO–P2O5–B2O3  and BaO–SnO–P2O5 ) with very low PEC values. Both compositions also showed fairly good water durabilities [1, 4], despite being phosphate glasses. However, the low temperature of crystallization still remains an issue when the phosphate glasses are reshaped into fiber form . In contrast, silicate glasses generally show greater crystallization temperatures. Therefore, we tried to develop zero-PEC tin silicate glasses as components of new optical current sensors. Although the properties and structure of the binary SnO–SiO2 glass system have been already studied [6, 7], the photoelastic characteristics have not been covered in the literature.
A systematic study of the PEC in binary SnO–SiO2 glasses has been reported by the Zwanziger’s group . This study predicted a zero-PEC (< 0. 05 × 10−12 Pa−1) composition, although it has not been prepared yet. Acquisition of a glass composition with zero PEC depends on accuracy of measurement and quality of glass that leads to less fluctuation of PEC value. The method we utilized has an advantage to determine precise retardation (PEC), since a heterodyne laser system with a coherent Zeeman laser has a capability with an error of PEC less than ± 0.01 × 10−12 Pa−1 . The drawback of the method is to choose a single wavelength (here 632.8 nm) for determining PEC. On the while, one of the advantages of Sénarmont method lies in choosing various wavelengths for retardation (PEC) in glass, although the accuracy of the measurement has been mentioned at about ± 10% . One of the difficulties in preparing those composition series lies in the glass making process . Regarding sample quality, most problematic point is residual thin crystalline film of SnO2 on the bulk glass. In present, optimized glass can be made under inert atmosphere with low partial pressure of oxygen during melting, as is mentioned below. Thus, suppressing the oxidation of Sn2+ ions while melting the batched raw materials (i.e., SnO and SiO2) is strongly required as these Sn2+ species have larger electron densities (i.e., metallic character) [9, 10], compared with undesired oxidized Sn4+ ions.
High polarizable Sn2+ ions play a nearly similar role as Pb2+ ions in tin silicate glasses . The polarizability of oxygen ions is influenced by the polarization of the cations surrounding the high polarizable Sn2+ ions. Therefore, the Sn2+ ions attract the electron clouds from oxygen ions, thereby deforming them. This effect may result in a spatially uniform response to the electric field generated by an incident light and to the suppression of the birefringence in case of atomic displacements of high polarizable Sn2+ and oxygen ions by macroscopically induced external forces.
According to the literature [8, 11], cations combining high polarizability (i.e., metallicity) and low coordination numbers with the nearest oxygen atoms are necessary for avoiding positive PEC values in some oxide glasses. This concept is clearly represented by the empirical term d/Nc. Thus, large d (i.e., shorter cation–oxygen bond length) and low Nc (i.e., coordination number with the nearest oxygen atoms) values account for PEC variations to a large extent. Our previous study on ZnO–SnO–P2O5–B2O3 glasses successfully supported this concept . In this letter, we report a detailed binary SnO–SiO2 composition forming a silicate network glass structure with zero PEC.
The glass samples were prepared by a conventional melt quenching method. The general formula was xSnO–(100−x)SnO (in molar ratio). The starting materials were commercial SnO (99.5%, KOJUNDO Chemical Laboratory Co., Ltd.) and SiO2 (99.99%, KOJUNDO Chemical Laboratory Co., Ltd.). These materials were weighed (5–10 g) and subsequently mixed using an agate mortar and a pestle under air. The samples placed in an alumina crucible provided with a lid were melted for 60 min at 1050–1250 °C in an alumina tube furnace under flowing argon with an O2 partial pressure lower than 10−3 atm to minimize oxidation of Sn2+ to Sn4+ in the melt. The melt was quenched in a crucible under cool air and immediately annealed at 470–550 °C (depending on their glass composition) to remove the residual thermal stress. These annealing temperatures were 10°C above the glass transition temperature (Tg) determined by differential thermal analysis. The amorphous nature of the resulting bulk glasses after removing the crystallized surface (identified as SnO2) was confirmed by X-ray diffraction analysis.
The serious points due to the surface crystallization are two; one is a light scattering on end surfaces of a fiber made by present SnO−SiO2 glass with zero PEC. The other is to diminish mechanical strength of the glass when it is elongated into a fiber under isothermal heating. In this study, we have found a condition of making glass without crystallization of inner bulk glass. Actually, the surface oxidation definitely occurs when the melt was quenched out of a furnace without inert gas purged.
The obtained glasses were cut and polished to disk shapes of ca. 10 or 20 mm diameter (D) and ca.10 mm length for PEC measurements. Disk samples with larger diameters were needed for examining glasses with low PEC (< 0. 5 × 10−12 Pa−1) in order to apply relatively high uniaxial loads up to ca. 30 kg. A frequency-stabilized transverse Zeeman He–Ne laser operating at 632.8 nm was used as a light source . The laser light that propagates both the half-wave plate and the linear polarizer has variations of the phase of intra-mode beat signal with the rotation angle of the half-wave plate. The rotation angle of the half-wave plate and that of the linear polarizer were synchronized each other with a ratio of 1 to 2 by each driving system. The phase difference between the fast and slow axes was maximized when the direction of a linearly polarized light was parallel to the fast axis. On the contrary, the difference was minimized in the direction perpendicular to the fast axis. The phase difference Δ was normalized by the wavelength of 632.8 nm, and was transformed to the retardation δ by the equation of δ [nm] = Δ [deg.] × 632.8/360 [nm/deg.]. The PEC (C) is given as C = πDδ/8P , where D is the diameter of sample, and P is the applied load. The positive or negative signs of PEC were determined by the direction of the detected fast axis for that of the external stress. The reproducibility of the PEC measurements for each glass was within ± 0.01 × 10−12 Pa−1 as a result of sample quality variations in our laboratory. This advantageous method , which leads to high precision PEC, is crucial for determining of the zero PEC composition. Combination of the method and relatively large sample mentioned above, which can provide with accurate measurement, successfully results in finding a composition with zero PEC. The optical transmittance spectra were measured using a 1-mm-thick plate.
3. Results and discussion
Figure 1 shows the transmittance spectra for binary SnO−SiO2 glasses. No marked absorption was observed in the near-infrared region up to a wavelength of 1600 nm. The SnO−SiO2 glass materials prepared herein showed adequate optical transmittance at a wavelength of 1550 nm, and are therefore designed to be operated in fiber-type current sensor devices. The optical transmittance of the samples was comparable to that of the 45PbO−55SiO2 glass sample used in the device . The absorption edges on those SnO−SiO2 glasses primary depended on the amount of SnO as the optical bandgap is determined by the 5s2–5p transition orbitals of Sn2+ ions . In addition, greater ligand fields formed by oxygen ions bound to silicon may be created around Sn2+ ions, this being strongly related to a short-range structure around Sn2+ ions in the glasses. The compositional dependence of the structure (e.g., coordination number of Sn2+) including the zero-PEC silicate glass composition should be revealed in detail in subsequent studies.
Figure 2 shows the variations of the average PEC values with the SnO concentration in SnO–SiO2 glasses measured at 632.8 nm. PEC decreased with the SnO concentration from + 3.53 × 10−12 Pa−1 to −2.37 × 10−12 Pa−1. The zero-PEC composition (i.e., ≤ 0.01 × 10−12 Pa−1) was reached at ca. x = 43.5 for the binary system. The error in the PEC determination in the present study was within ± 0.01 × 10−12 Pa−1, depending on the sample fluctuation. The reasons for deviation from the reported values were thought to be the different probing wavelengths, the method of PEC evaluation, and the sample fabrication procedure. The empirical term d/Nc in the binary SnO–SiO2 glass at nearly zero PEC was approximately 0.47 at x = 43.5 and using the parameters reported for SnO and SiO2 as proposed by Guignard et al .
PEC value measured with a wavelength of 632.8 nm can be giving access to useful information to predict in near infra-red region, since the wavelength locates close to the operating wavelength (1550 nm) of the present fiber type current sensor . According to the Sellmeier’s formula, which accounts for refractive index dispersion, the curvature between red (632.8 nm) and near infra-red region is mild than that between green (565 nm) and the infra-red region. If we can use a commercialized Zeeman laser with a wavelength of 1550 nm, it will be best system to estimate practicable PEC value, although some difficulties such as manipulating invisible light, and seeking for an appropriate half-wave compensator and filter for the infra-red light.
Figure 3 shows the Raman and Fourier transform infrared spectra of the sample under study. Both spectra showed a qualitatively consistency. Thus, the compositions were dominated by Q2 and Q1 units (in Qn, n denotes the number of bridging oxygens) . According to ref. 6, the fraction of Q2 and Q1 units obtained by analysis of the 29Si NMR spectrum of a 49.5SnO–50.5SiO2 glass were reported to be 38% and 24%, respectively. The compositional dependence of the structure in the present binary system including zero-PEC composition should be reported in subsequent studies.
Figure 4 shows schematic models of SnO–SiO2 glasses having a negative PEC. As shown in Fig. 2, PEC is very sensitive to composition (i.e., to the ligand structure around Sn2+ ions) and to the discrete silicate network. Species including longer silicate anions having different units were incorporated into the glasses, as indicated in Fig. 4 (a). The model in Fig. 4 (b) represents the atomic dislocation and the resultant extension of the electronic clouds belonging Sn2+ and O2- ions having high polarizability under a uniaxial loading on the glass. These electronic clouds were densified to the horizontal direction as a result of the dislocation of Sn2+ ions, meta-silicate (Q2), and pyro-silicate (Q1) units loosely bonded in the glass. As a result, the refractive index became greater in the horizontal axis direction (slow axis for light) perpendicular to the loading axis (fast axis).
The zero-PEC structure seems to present an intermediate situation between loosely bonded silicate units and reinforced Si–O–Si linkages around the Sn2+ ions. Although the coordination number around a Sn2+ ion in binary SnO–SiO2 glasses has been reported by Bent et al. , a quantitative experiment determining the fraction of silicate units and the coordination number of oxygens around a Sn2+ ion should be conducted for the present SnO–SiO2 composition having zero PEC. Fiber drawing conditions relevant to the thermal properties and a detailed composition dependence of the structure including the composition with zero PEC will be reported in subsequent studies.
In conclusion, a precise glass composition with zero PEC ( + 0.01 × 10−12 Pa−1) was revealed for binary SnO–SiO2 glasses free of hazardous elements. The composition exhibited high transparency (i.e., comparable to that of PbO–SiO2 glasses) for being utilized in optical fiber current sensors. The glass composition showed meta- and pyro-silicate units not tightly linked to Sn2+ ions in the silicate network. The structure of those shorter silicate units elucidated by Raman and FTIR analyses qualitatively suggested the presence of anisotropic electronic clouds after atomic displacement upon compression of the glass in the perpendicular direction.
We thank Prof. H. Toyota in Ehime University for use of micro-Raman scattering apparatus.
2. K. Kurosawa, “Development of fiber-optic current sensing technique and its applications in electric power systems,” Photonic Sensors 4(1), 12–20 (2014). [CrossRef]
3. K. Kurosawa, K. Yamashita, T. Sowa, and Y. Yamada, “Flexible fiber Faraday effect current sensor using flint glass fiber and reflection scheme,” IEICE Trans. Electron. E83-C, 326–330 (2000).
4. A. Saitoh, K. Nakata, G. Tricot, Y. Chen, N. Yamamoto, and H. Takebe, “Zero photoelastic and water durable ZnO–SnO–P2O5–B2O3 glasses,” APL Mater. 3(4), 046102 (2015). [CrossRef]
5. A. Saitoh, S. Anan, and H. Takebe, “Surface crystallization behavior during thermal processing of low-photoelastic ZnO–SnO–P2O5 glasses,” J. Mater. Sci. 52(4), 2192–2199 (2017). [CrossRef]
6. M. M. A. Karim, Doctoral Thesis, “A study of tin oxides in silicate based glasses,” University of Warwick, 1995.
7. J. F. Bent, A. C. Hannon, D. Holland, and M. M. A. Karim, “The structure of tin silicate glasses,” J. Non-Cryst. Solids 232–234, 300–308 (1998). [CrossRef]
8. M. Guignard, L. Albrecht, and W. Zwanziger, “Zero-stress optic glass without lead,” Chem. Mater. 19(2), 286–290 (2007). [CrossRef]
9. M. Tashiro, “The effects of the polarisation of constituent ions on the photoelastic birefringence of the glass,” J. Soc. Glass Technol. 40, 353T–362T (1956).
10. W. A. Weyl and E. C. Marboe, “The constitution of Glass,” (Wiley-Interscience, New York, 1964), Vol. 2.
11. V. Martin, U. Werner-Zwanziger, J. W. Zwanziger, and R. A. Dunlap, “Correlation of structure and photoelastic response in tin phosphate glass,” Int. J. Appl. Glass Sci. 2(4), 282–289 (2011). [CrossRef]
13. D. Ehrt, “Effect of OH-content on thermal and chemical properties of SnO–P2O5 glasses,” J. Non-Cryst. Solids 354(2-9), 546–552 (2008). [CrossRef]
14. P. McMillan, “Structural studies of silicate glasses and melts-applications and limitations of Raman spectroscopy,” Am. Mineral. 69, 622–644 (1984).