A flexible chalcogenide fiber bundle (FB) with a resolution as high as ~31 lp/mm has been fabricated for delivering thermal images of objects at room temperature. The FB is composed of ~200,000 single fibers with a Ge-As-Te-Se glass core 15 μm in diameter and a polyetherimide (PEI) cladding 16.8 μm in diameter. These Ge-As-Te-Se/PEI fibers show good transparency in the 3-12 μm spectral region. The fabricated FB presents a filling factor of ~72% and a crosstalk of ~1%. High-quality thermal images of a human hand were obtained through the FB, demonstrating good potential of the FB for longwave infrared imaging in the areas such as medicine, industry and defense.
© 2017 Optical Society of America
Infrared (IR) thermography has been widely used for medical applications since it can provide real-time imaging of temperature variations on a human body [1,2]. In an effort to obtain the heat distribution inside a human body, IR fiber bundles (FBs) were proposed to gather and deliver thermal images of the targeting tissues [3,4]. This makes it possible to monitor the tissue temperatures during ablation surgery, or even to achieve early detection of cancers in various internal organs. Currently, FBs developed for transmitting thermal images utilize polycrystalline silver halide (PSH) fibers [5–7]; hollow capillary (HC) fibers [4,8,9]; hollow-core anti-resonant (HCAR) fibers ; and chalcogenide glass (ChG) fibers [11–13]. PSH FBs consisting of 100-9000 individual fibers with a diameter of 20-250 μm have been fabricated by a multiple-extrusion method [6,14]. However, the extrusion procedure tends to cause nonuniform areas and different orientations in each bunch of fibers; and the FBs fabricated by this approach typically show a crosstalk of as high as 25% . HC FBs have been manufactured through a process involving the assembly of single HC fibers and the deposition of thin metal and dielectric layers on the inner surface of the glass capillaries . Unfortunately, HC fibers with a bore diameter of 100 μm showed transmission losses as high as tens of dB/m, and the losses are inversely proportional to the third power of the bore diameter [8,15]. In comparison, HCAR fibers are expected to show much lower losses in their transmission windows because of the anti-resonant guiding mechanism. Recently, a borosilicate glass HCAR FB composed of 245 individual fibers with a core diameter of 60 μm was fabricated by a stack-and-draw technique. The FB showed relatively low losses (around 15 dB/m) in the 3~4 μm wavelength range . Nevertheless, the loss increased dramatically when the wavelength exceeded 5 μm due to the strong absorption of the oxide glass. To date, the reported PSH, HC and HCAR FBs have a resolution not exceeding 6 lp/mm, which is too low for transmitting fine thermal images. In comparison, ChG FBs with a relatively high resolution of up to 10 lp/mm can be fabricated using the conventional method of ‘ribbon stacking’ (or layer winding) . More recently, an As2S3/PEI ChG FB with an ultra-high resolution of 45 lp/mm was achieved through a modified stack-and-draw approach, and high-quality thermal images were delivered using the FB . Nevertheless, that FB could only transmit light in the 1-7 μm. From the perspective of thermal imaging in clinical applications, FBs which can propagate longwave IR light up to 12 μm are desirable, because human tissues at room temperature show the most intense radiation in 8-12 μm range , and low-cost detectors (e.g. uncooled microbolometers) are available at this spectral region. In this paper, we report a flexible ChG FB based on optimized Ge-As-Te-Se (e.g. Ge10As30Te38Se22) glass, which has good transmittance in the 3-12 μm spectral region. This FB consists of about 200,000 individual fibers with a diameter of ~16.8 μm. It has a resolution of 31 lp/mm, and is capable of transmitting high-quality thermal images of human tissue.
2. Experimental procedures
2.1 Glass preparation
In this study, Ge-As-Te-Se glass was chosen as the core material because it shows good transparency in the longwave IR region, and has superior thermal stability against crystallization during fiber drawing . Although Se-free tellurite glasses (e.g. Ge-Te, Ge-As-Te, Ge-Ga-Te, etc.) have longer IR cut-off edges, they suffer from a strong tendency to crystallize and severe free-carrier absorption [18–20]. The Ge-As-Te-Se glass was prepared by melting mixtures of high purity constituent elements (5N Ge, and 6N As, Te and Se) in evacuated (<10−5 torr) and sealed low-OH (<2 ppm) quartz tubes in a rocking furnace. The material weighing and loading were conducted in a glove box with controlled water and oxygen concentrations of <0.1ppm. The mixtures were homogenized at 850°C for 12 hours. After that, the tubes containing the melts were quenched in water and the formed glasses were annealed to minimize inner stresses. In order to get high purity glasses, As and Se were pretreated under vacuum at 320°C and 240°C, respectively, for 2 h to remove the volatile oxide impurities. To further improve the glass purity, an elaborate dynamic distillation approach  was also applied. 200 ppm weight of ultra-dry GaCl3 (5N) and 20 ppm weight of Al (5N) were used as the C/H scavenging agent and oxygen getter, respectively. The details of this approach can be found in . The dimension of the glass rods so-obtained was about Ф15 mm × 100 mm.
2.2 Fiber bundle fabrication
A high-strength thermoplastic polymer, polyetherimide (PEI), was used as the cladding material, because it is thermally compatible with the Ge-As-Te-Se glass, and can also protect the fragile glass core from being broken during the whole fabrication process. The FB was fabricated by an improved stack-and-draw method as illustrated in Fig. 1. Fibers with a diameter of ~400 μm were firstly drawn from cylindrical preforms composed of a Ge-As-Te-Se glass core with a diameter of 15 mm and a PEI cladding with a diameter of 17 mm [Fig. 1(a)]. In the second step, the fibers were cut into 200 mm long segments, and about 800 segments were closely stacked in a hexagonal die with an inner side length of 6.5 mm [Fig. 1(b)]. The die was then put into a vacuum furnace for thermal consolidation of the inner materials leading to a hexagonal Ge-As-Te-Se/PEI column. The hexagonal column so-obtained was subsequently drawn into multi-fibers with a side length of ~280 μm [Fig. 1 (b)], which were then cut into segments with a length of 200 mm or 50mm. In the third step, about 250 multi-fiber segments were closely stacked in another hexagonal die with an inner side length of 4 mm [Fig. 1(c)]. In the last step, both ends of the stacked multi-fiber segments were rolled with protective PEI films and the rolled parts were carefully thermally-consolidated under vacuum. Thus, flexible FBs consisting of ~200,000 single fibers were formed [Fig. 1(d)].
The characteristic temperatures of the materials were measured at a heating rate of 10°C/min by a TA Q2000 differential scanning calorimeter (DSC, TA Instruments, New Castle, DE). A sample of about 10 mg was sealed into an aluminum pan for the measurement. The measurements of refractive indices of the materials were conducted on single-side polished discs using an IR-VASE ellipsometer (J. A. Woollam, Lincoln, NE) [22,23]. The transmission spectra of the fibers were recorded by a tensor 27 Fourier-transform infrared spectrophotometer (FTIR, Bruker, Ettlingen, Germany) equipped with an external fiber coupling unit and a liquid-nitrogen-cooled HgCdTe detector, and the losses of the fibers were determined by the cut-back method. The cross section of the FB was observed with an Axio scope A1 optical microscope (Zeiss group, Oberkochen, Germany). The performance of the FB for delivering thermal images was tested by a homemade imaging setup. The target was imaged on the input end of the FB through a ZnSe lens. Another ZnSe lens was used to collect the transmitted light from the output end and image it onto a FLIR T640 camera, which operates in the longwave IR region of 7.5 ~13 μm.
3. Results and discussions
By assessing the thermal stability against crystallization and fiber drawing temperatures of different compositions in Ge-As-Te-Se glass system, we found the composition Ge10As30Te38Se22 (GATSe), which has a glass transition temperature (Tg) of ~185 °C, did not show any trace of crystallization after multiple fiber drawings at 285-295°C. This makes the composition ideal for the fabrication of high-resolution FBs. The PEI which has a Tg of ~215 °C could be drawn into fibers in a wide temperature range of ~260-350 °C. The tests on fiber drawing using a preform with a Ge10As30Te38Se22 core and a PEI cladding showed that fibers with good optical quality could be drawn at ~290 °C. Figure 2 displays the attenuation of a fabricated GATSe/PEI fiber with a core diameter of 354 μm and a cladding diameter of 400 μm. For comparison, the attenuation of an As2S3/PEI fiber  and the black-body radiation spectrum at the temperature of 37 °C are also shown in the Figure. The fiber shows good transparency in the 3-12 μm spectral region, which overlaps with the wavelength range of the intense radiation from a black-body at 37 °C. The background loss is about 0.8 dB/m. The relatively high losses at 4.6 μm and 5.0 μm are due to the absorptions of Se-H and Ge-H impurities , respectively.
Figure 3(a) shows the cross section of the fabricated FB with a length of 50 mm. The individual fibers has a core diameter of ~15 μm, and the pitch between the cores is ~16.8 μm. The size of the individual fibers is comparable to that of the pixel (e.g. 17 μm × 17 μm) of commercially available high-resolution microbolometers. This makes the FB suitable for transmitting high-quality thermal images. The filling factor (or active area) of the FB is about 72%, which is quite high for optical FBs. Figure 3(b) displays the thermal image of a human hand captured by the FLIR T640 camera through the 50 mm long FB. The image is of high quality and clearly shows the fine features of the hand, indicating excellent performance of the FB for transmitting thermal images.
The resolution and crosstalk are key numerical parameters for characterizing the optical quality of FBs. The resolution can be determined from the modulation transfer function (MTF), which is typically measured using the knife-edge method [6, 25]. The crosstalk may be obtained by coupling light into a single fiber and measuring the intensity distribution of the transmitted light on the output end of the FB [16, 25]. Considering that the detector of FLIR T640 camera could be easily damaged by bright longwave IR light, we used mid-wave IR light for the tests because our Xenics InSb camera was capable of detecting bright light of 1-5 μm. In the measurements, 3.8 μm light generated from an optical parametric amplifier (OPA) was used. The measurement schemes are similar to those detailed in . Figure 4(a) shows the measured MTF of the fiber bundle. The FB presents a resolution of ~31 lp/mm, which is quite high for IR FBs. Figure 4(b) displays the image of the output after the 3.8 μm light propagates through a single fiber in the FB. The measured crosstalk is ~1%, which is quite low for optical FBs.
The resolution of a FB is related to the diameter of the single fibers and the filling factor of the FB. Closely packed single fibers with a smaller diameter generally result in a FB with a higher resolution. Compared with previously reported As2S3/PEI FB which has a filling factor of ~50%, in the fabrication of the GATSe/PEI FB, the stacked 800 segments were not rolled in protective PEI films. This leads to much smaller gaps between the multi-fibers in the FB, and therefore a remarkably higher filling factor (e.g. 72%). The GATSe/PEI FB shows a lower resolution than the As2S3/PEI FB, due to the fact that its individual fiber had a larger diameter. In fact, it is not necessary to further reduce the diameter of the single fibers in GATSe/PEI FB for higher resolution, because the size of a single fiber was already comparable to that of a pixel on the high-resolution microbolometer. For the individual fibers, the GATSe core glass and PEI cladding have refractive indices of ~2.88 and ~1.65, respectively. The ultra-high refractive index contrast makes the fibers possess excellent light confinement ability. In addition, PEI has strong absorptions in the mid- and longwave IR, preventing the light in the individual fiber from leaking into its neighbor ones. This may account for the observed low crosstalk of the FB.
The demonstrated GATSe/PEI FB had a relatively large cross section and a short length. However, the long multi-fiber which had a side length of 280 μm obtained after the second fabrication step can be considered to be a thin FB consisting of ~800 single fibers with a diameter of ~16.8 μm. This thin FB should have similar performance to the one described in this work, because it has almost the same single-fiber diameter and filling factor. This kind of thin FB may be more suitable for the delivery of thermal images in clinical applications because of its smaller size, better flexibility, as well as much lower cost.
It is worth mentioning that the single fibers in the FB were 16.8 µm in diameter and should have much higher losses than the ones that were measured in Fig. 2 whose diameter was 400 μm because of the enhanced interaction between the propagating light and the absorbing PEI. Unfortunately, our FTIR system was not able to measure the losses of the single fibers because the core size was too small to transmit detectable light. In order to estimate the attenuation level of the individual fibers, 10 μm light from an OPA was coupled into a single fiber from the 50 mm long FB, and a transmittance of ~30% was obtained. The loss of the single fiber could then be estimated to be ~60 dB/m after deducting the end-face reflections (~40%). Our measurements also showed that the 50 mm long FB had a total transmittance of ~21% at 10 μm. However, the transmittance of the 200 mm long FB was negligibly at 10 μm, and no detectable thermal images could be captured by the FLIR T640 camera through this FB. However, ultimately the FB may need to be ~1 m or longer for practical applications. To deliver reasonably clear thermal images, the FB would need to have a total transmittance of >10%, and consequently the losses of the single fibers need to be reduced to <8 dB/m if antireflection coatings are applied to the end faces of the FB. This attenuation level should be achievable by adding a thin Ge-As-Te-Se inner cladding with a lower refractive index between the core and the PEI . This work is still in progress.
A flexible GATSe/PEI FB was fabricated through an improved stack-and-draw method. The FB consisted of ~200,000 single fibers with a diameter of 16.8 μm. The fibers showed good transparency in the wavelength range of 3-12 μm. The fabricated FB had a high filling factor of ~72%, a high resolution of ~31 lp/mm, and a low crosstalk of ~1%. High-quality thermal images of a human hand were delivered through the FB, demonstrating its good performance for transmitting thermal images of objects at room temperature, and therefore indicating their good potential for longwave IR imaging in medical, industrial and defensive applications.
National Natural Science Foundation of China (NSFC) (51303072, 61575086, 61405080); Research Innovation Program for College Graduates of Jiangsu Province (KYZZ16_0470); Priority Academic Program Development of Jiangsu Higher Education Institutions; Jiangsu Collaborative Innovation Centre of Advanced Laser Technology and Emerging Industry;Xuzhou Science and Technology Project (KC16SG267); Australian Research Council through Centres of Excellence Program (CE110001018).
References and links
1. E. F. Ring and K. Ammer, “Infrared thermal imaging in medicine,” Physiol. Meas. 33(3), R33–R46 (2012). [PubMed]
2. B. B. Lahiri, S. Bagavathiappan, T. Jayakumar, and J. Philip, “Medical applications of infrared thermography: A review,” Infrared Phys. Technol. 55(4), 221–235 (2012).
3. I. Gannot, “Thermal imaging bundle-A potential tool to enhance minimally invasive medical procedures,” IEEE. Circ. Dev. Mag. 21(6), 28–33 (2005).
4. Y. Matsuura and K. Naito, “Flexible hollow optical fiber bundle for infrared thermal imaging,” Biomed. Opt. Express 2(1), 65–70 (2010). [PubMed]
5. I. Paiss and A. Katzir, “Thermal imaging by ordered bundles of silver-halide crystalline fibers,” Appl. Phys. Lett. 61(12), 1384–1386 (1992).
6. E. Rave, D. Shemesh, and A. Katzir, “Thermal imaging through ordered bundles of infrared-transmitting silver-halide fibers,” Appl. Phys. Lett. 76(14), 1795–1797 (2000).
7. Y. Lavi, A. Millo, and A. Katzir, “Thin ordered bundles of infrared-transmitting silver halide fibers,” Appl. Phys. Lett. 87(24), 241122 (2005).
8. U. Gal, J. Harrington, M. Ben-David, C. Bledt, N. Syzonenko, and I. Gannot, “Coherent hollow-core waveguide bundles for thermal imaging,” Appl. Opt. 49(25), 4700–4709 (2010). [PubMed]
9. C. Huang, S. Kino, T. Katagiri, and Y. Matsuura, “Remote Fourier transform-infrared spectral imaging system with hollow-optical fiber bundle,” Appl. Opt. 51(29), 6913–6916 (2012). [PubMed]
10. T. Kobayashi, T. Katagiri, and Y. Matsuura, “Multi-element hollow-core anti-resonant fiber for infrared thermal imaging,” Opt. Express 24(23), 26565–26574 (2016). [PubMed]
11. P. Klocek, M. Roth, and R. D. Rock, “Chalcogenide glass optical fibers and image bundles - properties and applications,” Opt. Eng. 26(2), 88–95 (1987).
12. A. R. Hilton, A. R. Hilton, J. McCord, W. S. Thompson, and R. A. Leblanc, “Infrared imaging with fiber optic bundles,” Proc. SPIE 5074, 849–854 (2003).
13. S. B. Mobley, B. Shaw, D. Gibson, V. Nguyen, R. Gattass, J. Sanghera, and I. Aggarwal, “IR imaging bundles for HWIL testing,” Proc. SPIE 8015, 801503 (2011).
14. E. Rave, L. Nagli, and A. Katzir, “Ordered bundles of infrared-transmitting AgClBr fibers: optical characterization of individual fibers,” Opt. Lett. 25(17), 1237–1239 (2000). [PubMed]
15. V. Gopal, J. A. Harrington, A. Goren, and I. Gannot, “Coherent hollow-core waveguide bundles for infrared imaging,” Opt. Eng. 43(5), 1195–1199 (2004).
16. B. Zhang, C. Zhai, S. Qi, W. Guo, Z. Yang, A. Yang, X. Gai, Y. Yu, R. Wang, D. Tang, G. Tao, and B. Luther-Davies, “High-resolution chalcogenide fiber bundles for infrared imaging,” Opt. Lett. 40(19), 4384–4387 (2015). [PubMed]
17. Z. Yang, T. Luo, S. Jiang, J. Geng, and P. Lucas, “Single-mode low-loss optical fibers for long-wave infrared transmission,” Opt. Lett. 35(20), 3360–3362 (2010). [PubMed]
18. B. Bureau, S. Danto, H. Ma, C. Boussard-Pledel, X. Zhang, and J. Lucas, “Tellurium based glasses: A ruthless glass to crystal competition,” Solid State Sci. 10(4), 427–433 (2008).
19. Z. Yang, O. Gulbiten, P. Lucas, T. Luo, and S. Jiang, “Long-wave infrared-transmitting optical fibers,” J. Am. Ceram. Soc. 94(6), 1761–1765 (2011).
20. S. Maurugeon, B. Bureau, C. Boussard-Pledel, A. Faber, P. Lucas, X. Zhang, and J. Lucas, “Selenium modified GeTe4 based glasses optical fibers for far-infrared sensing,” Opt. Mater. 33(4), 660–663 (2011).
21. B. Zhang, W. Guo, Y. Yu, C. Zhai, S. Qi, A. Yang, L. Li, Z. Yang, R. Wang, D. Tang, G. Tao, and B. Luther-Davies, “Low Loss, high NA chalcogenide glass fibers for broadband mid-infrared supercontinuum generation,” J. Am. Ceram. Soc. 98(5), 1389–1392 (2015).
22. Y. Yang, Z. Y. Yang, P. Lucas, Y. W. Wang, Z. J. Yang, A. P. Yang, B. Zhang, and H. Tao, “Composition dependence of physical and optical properties in Ge-As-S chalcogenide glasses,” J. Non-Cryst. Solids 440, 38–42 (2016).
23. J. A. Woollam, B. D. Johs, C. M. Herzinger, J. N. Hilfiker, R. A. Synowicki, and C. L. Bungay, “Overview of variable-angle spectroscopic ellipsometry (VASE): I. Basic theory and typical applications,” Proc. SPIE CR72, 3–28 (1999).
24. G. E. Snopatin, V. S. Shiryaev, V. G. Plotnichenko, E. M. Dianov, and M. F. Churbanov, “High-Purity Chalcogenide Glasses for Fiber Optics,” Inorg. Mater. 45(13), 1439–1460 (2009).
25. Y. Lavi, A. Millo, and A. Katzir, “Flexible ordered bundles of infrared transmitting silver-halide fibers: design, fabrication, and optical measurements,” Appl. Opt. 45(23), 5808–5814 (2006). [PubMed]