An important measuring technique under study for the DARWIN planet finding mission, is nulling interferometry, enabling the detection of the weak infrared emission lines of an orbiting planet. This technique requires a perfect wavefront of the light beams to be combined in the interferometer. By using a single mode waveguide before detection, wavefront errors are filtered and a virtually perfect plane wavefront is obtained. In this paper the results on the development and the optical characterisation of suitable infrared transmitting chalcogenide glasses and mid-IR guiding optical fibers are reported. Two different preform techniques for manufacturing core-cladding chalcogenide fibers are described. Two types of step index fibers, prepared with Te2As3Se5 chalcogenide glasses, offer single mode guidance at 10.6 µm.
© 2007 Optical Society of America
In this paper, recent results on the development of step index, single mode chalcogenide glass fibers are presented. Preceding the experimental part, the backgrounds and the specific aims of this fiber development for the envisaged space application, are outlined. For preparation of a single mode chalcogenide glass fiber, the main difficulties are the determination of the precise refractive index value of the core and the clad glasses as well as the combination of these two glasses in an optical configuration, leading to single mode propagation. In order to reach this goal, refractive index values were measured as a function of chemical composition using a surface reflection method.
To adjust the core/clad diameter ratio and to ensure a suitable interface between the core and the clad glasses, two method were used for manufacturing step index fibers. The first technique is an internal built-in-casting method that has been described before . The second method is a modified rod-in-tube technique.
To characterize the propagation modes and to verify that the characteristics of the light guide are consistent with the specifications set by the space application, far field experiments were conducted using a CO2 laser operating at 10.6µm as a light source.
2.1 IR waveguides for DARWIN mission
Nulling interferometry is an important measuring technique for future space programs, such as the DARWIN mission of European Space Agency (ESA) and the Terrestrial Planet Finder (TPF) of National Aeronautic Space Agency (NASA) [2, 3]. The main goal of these missions is to identify terrestrial planets, orbiting around nearby stars and having an earth-like atmosphere, so possibly supporting life. The DARWIN system, with capability for imaging and spectroscopy science, operating in the thermal infrared spectral region requires that wavefront errors be reduced to a very high degree in order to achieve the required nulling quality. DARWIN will consist of a nulling interferometer, combining light from several telescopes that are phase shifted from each others. When the beams of light are combined, they form a destructive interference pattern along the optical axis of the system and a constructive interference displaced by a small angle. This technique results in the light from the bright star being cancelled out, leaving only light from the planets around the star. If the signal emerging from the DARWIN beam combiner null output is to be passed directly to the detector (without wavefront filtering), a very high wavefront quality (>λ/6000) is required, which renders the entire optical system design and manufacturing unfeasible. Such a high wavefront quality can only be achieved with adequate wavefront filtering measures like a single mode waveguide. Indeed, the use of a single mode waveguide is an attractive way to eliminate amplitude and phase distortion of the beam. The role of such a guide is well detailed in ref . Since the main infrared signatures of the relevant atmospheric components like CO2, O3 and water vapour are all in the mid IR range, from 4–20 µm, the single mode waveguides must be transparent in this spectral range.
Thus, the role of the single mode waveguide for DARWIN is to filter infrared signals in the whole operating range from 4–20 µm with the lowest optical loss and highest achievable coupling efficiency. One practical requirement is that the mode field diameter (MFD) has to be much smaller than the cladding diameter of the waveguide to ensure the confinement of the mode. The MFD is defined as the diameter where the mode field intensity is dropped to 1/e2 of the maximum value. A single mode optical waveguide, based on a step-index configuration, is designed to be single mode for wavelengths above a certain cut-off wavelength value. Assuming a fix focal lens coupling optics, the coupling efficiency has a maximum at a certain wavelength (for example the cut-off wavelength) and drops very rapidly for the rest of the operational wavelength range. This constraint makes the development and operation of a single mode waveguide very challenging over such broad wavelength range. Another challenge is the low level of maturity of materials and manufacturing processes in the mid-infrared spectral region.
The requirements for the single mode waveguide as specified by ESA are presented in the table below.
2.2 Single mode glass fibers for the mid IR
Different types of single mode guides are being studied to meet the DARWIN requirements: microstructured fibers , rib waveguides , hollow waveguides  and conventional step-index fibers which are the object of the present work. The waveguide needs to transmit light in a broad infrared range. Two families of materials have a high potential to be used for the fabrication of such a device: infrared glasses or silver halide polycristals.
Infrared step-index fibers based on polycrystalline materials have been developed in the framework of the Terrestrial Planet Finder program, TPF, funded by the NASA in one hand [7,8], and in the framework of the Darwin ESA program on the other hand . These authors have shown that their fibers behave as single mode guides at 10,6 µm [8,9]. In , the optical losses were evaluated to be in the range 20–30 dB.m-1. More recently, another group succeeded in designing microstructured fiber made of silver halide polycristals . This device behaves also as a single mode guide with much lower losses: around 2 dB.m-1 measured by the cut-back method.
Infrared glasses showing a spectral transmission up to wavelengths beyond 15 µm include halide glasses which are known to be highly hydroscopic [11, 12], and the chalcogenide glass family [13–18]. In the vast chalcogenide glass family, some specific compositions, especially those based on As, S, Se, exhibit excellent glass forming ability and good chemical and mechanical properties that make them suitable for fiber drawing. Most of the attention has been paid to elaborate optical fibers for the telecommunication industry. These fibers usually cover the near infrared region (<5 µm) [19–21] and/or are not single mode , making them unsuitable for the DARWIN space application.
In order to exhibit the higher transparency in the mid infrared range, low phonon glasses are necessary. Consequently, the best compromise is a glass composition containing selenium combined to a maximum of tellurium, while keeping a strong resistance toward crystallisation. The selected glass composition, transparent in the lower wavelength part (4–12 µm) of the DARWIN spectral region, is the so-called TAS glass having the atomic composition Te2As3Se5. From this composition several bare (‘mono-index’) fibers have already been prepared. The typical attenuation curve for this multimode waveguide exhibits a transmission window from 2 to 12 µm with attenuation in the range of 2 dB.m-1 at 10.6 µm . Thus, a single mode TAS glass fiber could be envisaged to be used in the lower DARWIN wavelength range (<12µm). The feasibility of such a single mode fiber has been demonstrated  and technical notes have been written on the stage development of this device [25–27]. In the following, the authors propose to gather all the relevant results that lead to the first single mode optical fiber made from glass, for which single mode propagation at 10.6 µm was demonstrated.
Three glass rods composed of tellurium, arsenic and selenium elements were prepared to determine the effect of the Se substitution by the Te on the refractive index. The three different atomic compositions are Te20As30Se50, Te23As30Se47 and Te25As30Se45. One of the key operations in the glass synthesis is in the starting material purification. The synthesis and purification steps were already described and we invite interested people to read [15,23] for more information. In order to measure the refractive index, the glasses were cut along the rod axes in order to provide a suitable glass surface which will be optically polished.
3.2 Fibers elaboration
Step index core-clad fibers were manufactured using different preforms techniques, based on the modified well-known rod-in-tube perform method.
3.2.1 Internal built-in-casting method (IBCM)
By the IBCM method the following two operations are carried out in a closed silica vessel maintained under vacuum:
Firstly, the preparation of a clad tube by emptying the centre of the molten preform, which is fluid, while the exterior is still solid, Secondly, the filling of the core part of this preform by pouring the fluid core glass into the tube.
The details of this method are reported in [1, 28]. This original method guarantees a good interface between the two glasses but is delicate to implement, especially in terms of reproducibility. The typical final product obtained by this method is a preform of 10 cm length having a core diameter of about 2.5 to 3 mm, with an external diameter of 10 mm. This first step needs to be followed by several rod-in-tube operations as described below in order to reach the appropriate core to clad diameter ratio for a single mode fiber.
3.2.2 Rod-in-tube vacuum method (RTVM)
Another way to realize the first step preform preparation is to use the classical rod-in-tube method by collapsing the core and the clad glasses under vacuum. The cladding tube is prepared by rotational casting in order to reach an internal diameter of 2 mm and an external diameter of 10 mm. The 2 mm diameter rod to be inserted is obtained by drawing a larger 10 mm rod synthesized in a silica tube. During the stretching, an under-pressure is maintained at the interface between the core rod and the cladding tube in order to optimize the quality of the interface . This operation leads to a core/clad preform reduced to a 2 mm diameter stick. This preform is used as new rod to be inserted into a clad tube obtained by rotational casting. This operation is repeated in order to obtain the good core to clad ratio for single mode propagation at around 10 µm wavelength which means an internal diameter roughly equal to 20 µm for an external diameter equal to about 500 µm.
3.3 Optical measurements
3.3.1 Refractive index measurement
The refractive indices n(λ) were determined from the single surface reflection coefficient R(λ) according to :
in which nair=1.00030±0.0003 ,
The samples used for the reflection measurements were obtained by cutting a rod (20 mm diameter and 30 mm length) lengthwise and polishing the front flat surface.
The measurements in the IR range beyond 2.5 µm were performed with a Perkin Elmer 883 double beam, single monochromator IR spectrophotometer. Reflectance measurements were performed with the help of a specular reflectance accessory that is placed in the sample beam. The angle of incidence is 10° and the illuminated area on the sample is approximately 4×15 mm2.
Only the beam reflected specularly at the front surface is detected by the IR spectrophotometer, since the back surface of the sample has a semi-cylindrical shape, which will diverge the light reflected at the back surface. In the UV/Vis/NIR range (below 2.5 µm), reflectance measurements on these samples were limited to wavelengths below the UV absorption edge since in the transparent range the integrating sphere captures part of the back reflectance of the cylindrical samples. A smooth curve for n was obtained by fitting a 3-term Sellmeier equation:
3.3.2 Optical characterization by far field intensity distribution measurements
To characterize the mode profile of transmitted light through the fiber samples, 2D far field intensity (FFI) distribution measurements were carried out. The set-up used for FFI measurements at 10.6 µm wavelength basically consists of a CO2 laser and a 2D-array micro bolometer IR CCD camera. A pinhole of 20 µm diameter in front of the fiber sample ensures that the laser light is coupled mainly into the core of the fiber. The light transmitted by the fiber is analysed by the IR camera.
Several fiber samples have been coated with an absorbing Gallium coating to eliminate the propagation of higher order optical modes in the cladding (cladding modes). For this, Gallium is melted in a holder at a temperature of 55 °C. Then the TAS fiber is pulled through the molten Gallium several times to obtain a homogeneous coating. The thickness of the coating varies typically between 15 and 45 µm. TAS fiber samples up to a length of 50 cm have been treated successfully using this method.
4.1 Transmission of the glasses
Figure 1 demonstrates the necessity of purifying the starting materials by a two steps purification procedure. Each purification step corresponds to a removal of selenium oxide and arsenic oxide, and consequently a non-negligible quantity of Se and As is lost. One can estimate that selenium and arsenic deficiency is between 1% and 2% which makes it necessary to adapt the weight of the starting elements.
By using two purification steps, the bands around 12.7 µm and 15.4 µm (probably related to asymmetric and symmetric stretching vibration of the As-O-As bridge ) have been removed successfully, leaving only the broad multiphonon vibration band around 14.3 µm due to the 1st and 2nd overtones of two stretching modes in the arsenic-selenium glass matrix .
This band limits the usable spectral transmission window of Te-As-Se glass waveguides of a length of 10 cm or longer to wavelengths below 12 µm. For shorter waveguides of a few cm or less, the usable transmission window is limited to about 16 µm due to a strong absorption band peaking at 21 µm (of which only the tail is visible in Fig. 1). This band is assigned to the two-phonon modes of the As-Se bonds .
4.2 refractive index measurements
The dispersion curves (refractive index versus wavelength) of three different Te-As-Se glass compositions were determined from the reflection measurements. The results in Fig. 2 show the raw measurement data (average of 6 measurements) according to Eq. (1) and the smooth curves that were obtained by fitting with Eq. (2). Data obtained in the non-transparent region (<1.1 µm not shown in Fig. 2) was also included in the fit. The estimated uncertainty in these refractive indices is ±0.5%. The raw data is clearly distorted by systematic errors. The “jump” in the spectra as shown at 5 µm is generally an indicator for misalignment and beam distortion, for which the effect depends on the sensitivity of the instrument. At 5 µm our instrument inhibits a grating change that is accompanied by a change in sensitivity. Sudden changes in the refractive index spectra are also shown at the positions coinciding with the absorption bands at wavelengths 12.74 µm and 15.43 µm. However, judging from the extinction coefficients of these bands, their corresponding oscillators are not strong enough to cause such a significant change in the refractive index. These changes must therefore be caused by systematic errors in the reflection measurement. In spite of these systematic errors, the refractive index values for the glass containing 20% of Te agree well (within 0.2%) with data obtained for the same glass composition by Aio, Efimov and Kokorina, using the more accurate classical method of measuring the angle of minimum deviation in a prism .
From the results it is clear that by substituting Te for Se the refractive index of the glass can be modified in a controlled way. On the basis of this finding, the two compositions selected to manufacture the single mode fibers are Te20.2As30Se49.8 for the core and Te20As30Se50 for the clad, ensuring a Δn between core and clad equal to 3.10-3 in the working wavelength range.
4.3 Optical characterization of TAS fibers
The first step-index fiber was fabricated using the IBC Method described in the 3.2.1 section. The final core radius is equal to 11 µm and the cladding thickness 250 µm. Figure 3 shows the far field intensity distribution of a 23 cm long TAS fiber sample visualized in a 2D and 3D perspective. The double index fiber is properly coated with a homogeneous Ga-layer over the entire length. Application of the Ga coating aims at suppressing the transmission of cladding/leaking modes. This effect has been shown and illustrated in . Thus, the transmitted signal has only one remaining peak with a Gaussian shape, indicating single mode operation of the fiber. In particular, no large contribution from higher-order modes is visible. However, it is clear in Fig. 3, that the output intensity is not exactly circular, but rather elliptical. This is due to the geometry of the core which slightly deviates from a perfect circular shape. Furthermore, the attenuation of this single mode fiber was measured by the cut-back method using the CO2 laser at 10.6 µm. The attenuation at this wavelength was found to be about 0.1 dB.cm-1.
The second step-index fiber was manufactured with the RTV Method described in 3.2.2. Its core and clad radii are similar as for the IBCM fiber. From our experience, it appears that the RTV method provides better control of the final fiber quality parameters, including a smoother interface between core and cladding and a good circularity of the core. Figure 4 shows the far field intensity distribution of a 36 cm long TAS step index fiber sample, coated with an absorbing Ga-coating over the full length. The output intensity exhibits a Gaussian shape, characteristic of single mode propagation. Moreover, this fiber exhibits a circular intensity distribution. Once again, the propagation loss, evaluated by the cut-back method, is equal to about 0.1 dB.cm-1 as for the IBCM fiber.
In order to evaluate whether a guide exhibits single mode propagation, the so-called normalized waveguide frequency V has to be calculated according to the following formula:
where ρ is the core radius and λ the working wavelength. It is well-known that a guide becomes single mode for V<2.405. The TAS fiber characteristics are summarized in table 2. Knowing that the numerical aperture is about 0.13, the normalized waveguide frequency V=9.16/λ. For wavelengths higher than the cut-off wavelength λc=3.7 µm, step index TAS glass fibers are expected to be single mode. As depicted in Figs. 3 and 4, this result has been verified experimentally at 10.6 µm.
The ideal broadband single mode fiber for the DARWIN mission should have a Mode Field Diameter (MFD) which is proportional to the wavelength because the diameter of the focusing spot is linear with the wavelength. Figure 5 shows that for the TAS glass single mode fibers manufactured in this work, no significant discrepancy appears between the ideal MFD and the calculated MFD from 4 to 10 µm, ensuring a good coupling efficiency on this range. For longer wavelengths, the mode field stretches more and more into the cladding, but the fiber does not transmit more IR light.
Furthermore, in order to check its space compatibility the TAS glass has also been exposed to a relevant space environment, while its optical performance has been monitored before and after testing. It has been found that a γ-ray radiation dose of 0.1 kGy and a cooling procedure to cryogenic temperatures do not influence the IR transmission of TAS glasses.
In conclusion, a step-index fiber based on TAS glass appears to be a good candidate for a wavefront filter operating between 4 and 12 µm for the Darwin mission. This step-index fiber exhibits single mode propagation at 10.6 µm with an adequate coupling efficiency. The optical losses were measured to be around 0.1 dB.cm-1, independently of the preparation methods (IBCM or RTVM). This result has to be improved further in view of the ESA DARWIN requirement of maximum 1.5 dB loss for each filtering device, including the Fresnel (reflection) losses. Nevertheless, these experimentally observed attenuation values are promising compared to the loss values of around 10 dB.cm-1 at 10.6 µm in TAS glass rib waveguides, reported in .
The single mode TAS fibers have to be considered as a first step towards realization of the required wavefront filters for ESA’s DARWIN mission, since they cover only the lower part of the envisaged wavelength range, enabling detection of H2O and O3. For wavelengths beyond 12 µm, involving the CO2 absorption band at 14 µm, it will be necessary to use other materials. Promising materials are new glasses based on tellurium, recently discovered in our laboratories. Indeed, these new vitreous materials exhibit a transparency window extending from 6 to 25 µm [17–18].
This work was supported by the European Space Agency thanks to an ESTEC contract 18831/05/NL/HB.
References and links
1. D. Le Coq, C. Boussard-Pledel, G. Fonteneau, T. Pain, B. Bureau, and J.-L. Adam, “A new method of preform elaboration for chalcogenide fibers,” Glass Technol. 44, 132 (2003).
2. R. N. Bracewell, “Detecting nonsolar planets by spinning infrared interferometer,” Nature 274, 780–781 (1978). [CrossRef]
3. J. R. P. Angel, A. Y. S. Cheng, and N. J. Woolf, “A space telescope for infrared spectroscopy of earth-like planets,” Nature 322, 341–434 (1978). [CrossRef]
4. J. C. Flanagan, D. J. Richardson, M. J. Foster, and I. Bakalski, “Microstructured fibers for broadband wavefront filtering in the mid-IR,” Opt. Express 14, 11773–11786 (2007). [CrossRef]
5. C Vigreux-Bercovici, E. Bonhomme, A. Pradel, J. E. Broquin, L. Labadie, and P. Kern, “Transmission measurement at 10.6 µm of Te2As3Se5 rib waveguides on As2S3 substrate,” Appl. Phys. Lett. 90, 1–3 (2007). [CrossRef]
6. P. Labeye, J. E. Broquin, P. Kern, P. Noel, P. Saguet, L. Labadie, C. Ruilier, and V. Kirshner, “Infrared single-mode hollow conductive waveguides for stellar interferometry,” Proc SPIE 6123, 1–9 (2006).
7. B. Z. Dekel and A. Katzir “Graded index silver chlorobromide fibers for mid-infrared,” Appl. Opt. 44, 3343–3348 (2004). [CrossRef]
8. S. Shalem, A. Tsun, E. Rave, A. Millo, L. Nagli, and A. Katzir, “Silver halide single mode fibers for the middle infrared,” Appl. Phys. Lett. 87, 091103 (2005). [CrossRef]
9. R. Flatscher, O. Wallner, V. Artjushenko, and J. Pereira do Carmo, “Manufacturing of Chalcogenide and Silver-Halide Single-Mode Fibres for Modal Wavefront Filtering for Darwin,” Proc. ‘6th Internat. Conf. on Space Optics’ ESTEC , 27–30 (2006).
10. L. N. Butvina, O. V. Sereda, E. M. Dianov, N. V. Lichkova, and V. N. Zagorodnev, “Single-mode microstructured optical fiber for the middle infrared,” Opt. Lett. 32, 334–336 (2007). [CrossRef] [PubMed]
11. J. LucasR. W. Cahn, P. Haasen, E. J. Kramer, and J. Zarsiky, eds.,Halide glassesMaterials Sciences and Technology (Wiley-VCH, New York1991), Vol. 9, pp. 457–488.
12. B. Bureau and J. L. AdamG Meyer, D Nauman, and L Weseman, eds.,Non- oxide optical glasses: properties, structures and applicationsInorg. Chem. Highlights (Wiley-VCH, Weiheim2005) Vol. 2, pp. 365–392.
13. S. R. ElliotR. W. Cahn, P. Haasen, E. J. Kramer, and J. Zarsiky, eds.,Chalcogenide GlassesMaterials Science and Technology (Wiley-VCH, New York1991) Vol. 9, pp. 377–448.
14. X. H. Zhang, H. L. Ma, J. L. Adam, J. Lucas, G. Chen, and D. Zhao, “Thermal and optical properties of the Ga-Ge-Sb-Se glasses,” Mat. Res. Bull. 40, 1816–1821 (2005). [CrossRef]
15. V. S. Shiryaev, M. F. Churbanov, E. M. Dianov, V. G. Plotnichenko, J. L. Adam, and J. Lucas, “Recent progress in preparation of chalcogenide As-Se-Te glasses with low impurity content,” J. Opt. and Adv. Mat. 7, 1773–1779 (2005).
16. B. B. Harbison, C. I. Merzbacher, and I. D. Aggarwal “Preparation and properties of BaS-Ga2S3-GeS2 glasses,” J. Non-Cryst. Solids 213&214, 16–21 (1997). [CrossRef]
17. S. Danto, P. Houizot, C. Boussard-Pledel, X-H Zhang, F Smektala, and J Lucas “A family of far-infrared transmitting glasses in the Ga-Ge-Te system for space applications,” Adv. Funct. Mater. 16, 1847–1852 (2006). [CrossRef]
18. A. Wilhelm, C. Boussard-Plédel, P. Lucas, M. Riley, B. Bureau, and J Lucas “New tellurium based glasses for use in bio-sensing applications,” Proc SPIE 6433, 1–8 (2007).
20. C. FloreaM. BashkanskyZ. DuttonJ. SangheraP. PurezaI. Aggarwal “Stimulated Brilloiun scattering in single mode As2S3 and As2Se3 chalcogenide fibers,” Opt. Express 14, 12063–12070 (2006). [CrossRef] [PubMed]
21. P. Houizot, F. Smektala, V. Couderc, J. Troles, and L. Grossard, “Selenide glass single mode optical for nonlinear optics,” Opt. Mat. 29, 651–656 (2007). [CrossRef]
22. T. Kanamori, Y. Terunuma, S. Takahashi, and T. Miyashita “Chalcogenide glass fibers for mid-infrared transmission,” J. of Lightwave Technol. 2, 607–613 (1984). [CrossRef]
23. S. Hocdé, C. Boussard-Plédel, G. Fonteneau, and J. Lucas “Chalcogens based glasses for IR fiber chemical sensors,” Solid State Sci. 3, 279–284 (2001). [CrossRef]
24. V. S. Shiryaev, C. Boussard-Plédel, P. Houizot, T. Jouan, J.-L. Adam, and J. Lucas “Single-mode infrared fibers based on TeAsSe glass system,” Mat. Sci. and Eng. B 127, 2–3, 138–143 (2006). [CrossRef]
25. L. K. Cheng, A. J. Faber, W. Gielesen, C. Boussard-Pledel, P. Houizot, J. Lucas, and J. Pereira Do Carmo “Test results of the infrared single-mode fiber for the DARWIN mission,” Proc. SPIE 5905, 59051F (2005). [CrossRef]
26. L. K. Cheng, A. J. Faber, W. Gielesen, J. Lucas, C. Boussard-Plédel, P. Houizot, and J. Pereira do Carmo “Development of broadband infrared single-mode fibers for the DARWIN mission,” Proc. SPIE 6268, 62682F (2006). [CrossRef]
27. A. J. Faber, L. K. Cheng, W. L. M. Gielesen, C. Boussard-Plédel, P. Houizot, S. Danto, J. Lucas, and J. Pereira Do Carmo “Single Mode Chalcogenide Glass Fiber as Wavefront Filter for the Darwin Planet Finding Mission,” Proc. ‘6th Internat. Conf. on Space Optics’ ESTEC , 27–30 (2006).
28. D. Le Coq, C. Boussard-Plédel, G. Fonteneau, T. Pain, B. Bureau, and J. L. Adam “Chalcogenide double index fibers: fabrication, design and applications as chemical sensor,” Mat. Res. Bull. 38, 1745–1754 (2003). [CrossRef]
29. P. A. Van Nijnatten, “Accurate measurement of absorption spectra and refractive index of glass by spectrophotometry,” Glastech.Ber. Glass Sci. Technol. 77C, 136–148 (2004).
30. R. C. Weast, Handbook of Chemistry and Physics, 56th edition, (CRC Press, 1975).
31. M. F. Churbanov, I. V. Scripachev, G. E. Snopatin, V. S. Shiryaev, and V. G. Plotnichenko “High purity glasses based on arsenic chalcogenides,” J. Opt. Adv. Mat. 3, 341–349 (2001).
32. C. T. Moynihan, P. B. Macedo, M. S. Maklad, R. K. Mohr, and R. E. Howard “Intrinsic and impurity infrared absorption in diarsenic triselenide glass,” J. Non-Cryst. Solids 17, 369–38 (1975). [CrossRef]
33. L. G. Aio, A. M. Efimov, and V. F. Kokorina, “Refractive index of chalcogenide glasses over a wide range of compositions,” J. Non-Cryst. Solids 27, 299–307 (1978). [CrossRef]