Open Access
Add to BibSonomyAbstract
A bundle composed of 245 anti-resonant glass hollow optical fibers with a total diameter of 1 mm and fiber core diameter of 60 μm is fabricated for endoscopic infrared-thermal imaging. The bundle fiber shows low losses in the wavelength range of 3 to 4 μm owing to the anti-resonant effect of the thin glass wall. An image resolution of around 420 μm with a field-of-view of 3-mm diameter is obtained although crosstalk between adjacent fibers is observed. The experimental results of an imaging system using the fiber bundle with a half-ball lens at the distal end, which can be inserted into a working channel of endoscopes, are also shown.
© 2016 Optical Society of America
1. Introduction
Infrared thermal imaging is commonly utilized in medical applications since it is non-invasive and can provide real-time imaging of the heat distribution of the surface of a human body [1,2]. Infrared thermography has been already used in early detection of breast cancer and other malignant tumors [3,4] that observes slight temperature differences between normal tissue and pre-cancerous tissue that shows higher temperature because of higher blood vessel activity. Recently, endoscopic infrared thermography has been strongly desired because it enables detection of inflammation on the inner surface of internal organs that can lead to an early non-invasive diagnosis of cancer. In these applications, detection of early cancer tissue that is smaller than 10 mm and the temperature difference of around 1°C is required. Endoscopic thermal imaging is also desired in applications of laser hyperthermia to monitor the laser radiation effect. In this application, observation of laser radiation spots that are several mm in size where temperature rises 1-2°C from body temperature [5]. Although a delivery medium for infrared images, such as an optical fiber bundle, is essential for endoscopic uses, a feasible delivery medium has not appeared because the delivery medium has to transmit mid-infrared light emitted from the subject with body temperatures as black body radiation. We have developed a bundle of hollow optical fibers that transmits mid-infrared light with low losses [6,7]. However, to fabricate this bundled hollow optical fiber, complicated processes involving deposition of thin metal and dielectric layers on the inside of the glass capillaries were necessary. As a result, the diameter of each fiber was larger than 200 µm, and this limited the resolution of the imaging fiber bundle.
Here, we propose a multi-element hollow-core fiber for infrared thermal imaging that confines the light to each core of the fibers by using the anti-resonant effect [8,9]. The hollow-core anti-resonant fiber is a configuration of photonic crystal fibers [10, 11], and their feasibility for transmitting mid-infrared light has been shown [12–14]. The simplest guiding mechanism of this type of fiber is basically explained by a single tube model consisting of a thin dielectric layer whose thickness is comparable to the wavelength of the guiding light [15,16]. As far as we know, Miyagi et al originally proposed the fiber as a “tube-leaky waveguide” in 1980 for transmitting submillimeter wave and mid-infrared light [17,18]. However, this type of single tube, which consists of an anti-resonant fiber, had not been put into practice because the fabrication of a self-sustainable structure was difficult due to the very thin fiber wall [19,20]. Recently, anti-resonant hollow optical fibers made of thin glass walls have become feasible because the thin glass walls are supported by other glass walls in the microstructured fibers. In this paper, we propose a bundle of anti-resonant hollow optical fibers for infrared imaging. In the proposed bundled fiber, each tubular fiber is supported by adjacent fibers with the same dimensions.
2. Design
The structure of the proposed bundled anti-resonant hollow optical fiber is shown in Fig. 1. The bundle is made of a glass material whose absorption coefficient is reasonably low in the wavelength region of transmitted light, and it has a honeycomb structure to maximize the effective area for transmitting light. Unlike other types of photonic crystal fibers where the light is confined to only the central core by using the anti-resonant effect of both the glass wall and the surrounding air layers [8–11,16], the anti-resonant condition is satisfied only in the glass wall in the proposed hollow optical fiber. This is because in the structure shown in Fig. 1, the core diameter of each airy core is much larger than the wavelength of the guiding light [15]; therefore, light is confined to every core of the bundle.
When the optical thickness of the glass wall of each fiber is an odd multiple of quarter wavelengths, from the viewpoint of ray optics, the phases of the two rays reflected from the front and back sides of the glass wall match, and as a result, reflection at the inner surface of the fiber is enhanced. This is called the anti-resonant effect, and in this condition, light is confined to the airy core, and the HE11 mode becomes the fundamental propagation mode like other conventional fibers. Although the oversized fiber supports higher order modes as well, the HE11 mode is dominant because its transmission loss is much lower than those of other high order modes. Figure 2 shows the calculated reflectance of the glass as a function of wall thickness at the incident angles of light that correspond to the propagation angles of each mode in the fiber with a core diameter of 60 μm [19]. In this calculation, the wavelength was set to 4.5 µm by considering the wavelength range of the camera used in the experiment and the mean of the power reflectance of the TE and TM waves were used for the HE11 mode. As a glass material, we chose borosilicate glass because of the lower softening temperature (∼900°C) than that of silica glass (1650°C). Owing to the low melting point, a small electric furnace without water cooling can be used for glass drawing although the larger extinction coefficient of borosilicate glass may cause higher transmission loss. We assumed the complex refractive index of borosilicate glass to be n-jk = 1.44-j*0.0047 at a wavelength of 4.5 μm based on our measurement with bulk glass.
Fig. 2 Calculated reflectance of borosilicate glass as a function of wall thickness for incident angles that correspond to the propagation angles of low order modes in a hollow-optical fiber with an inner diameter of 60 μm. Dotted line is theoretical reflectance of silver.
We found that the reflectance for the HE11 and the TE01 modes are almost the same and they are much higher than that of the TM01 mode. By considering difficulty of beam coupling to the TE01 mode having a donut-shape power profile, the HE11 seems to become dominant in the fiber. We also found that the difference in the reflectance between borosilicate glass and silica glass with extinction coefficient of 5.5*10−4 at the 4.5-μm wavelength is smaller than 0.003 at the thickness of 1.0 μm, and therefore, we confirmed that the loss increase by using borosilicate glass in place of silica glass will be kept low. Under the same condition, the reflectance of metallic silver was calculated to be 0.92 as shown by the dotted line in Fig. 2 assuming the complex index of silver to be n-jk = 2.77-j*32.4 [21]. This means that, in some conditions where the reflectance of the glass wall is higher than 0.92, the transmission loss of the tubular glass fiber becomes lower than that of conventional hollow optical fibers composed of glass capillary tubes whose inner walls are coated with a silver layer. In addition, we confirmed that the effect of the material absorption slightly increases when the film thickness is large; therefore, we set the glass wall thickness to be in the range of 0.6 to 1.0 µm in the fabrication described in the next section.
3. Fabrication
Bundles of anti-resonant hollow optical fibers were fabricated by the so-called stack-and-draw technique [22]. The fabrication process is schematically shown in Fig. 3. We firstly prepared 150-mm long, glass capillaries with inner and outer diameters of 0.9 mm and 1.0 mm, and sealed one of the ends in advance. Then, hundreds of capillary tubes were stacked and inserted into a large diameter (~20 mm) glass tube (Fig. 3(a)). When drawing this preform into the shape of a bundled fiber, the gap between the capillaries was depressurized from the upper end by the vacuum pump, and the bores of the capillaries were pressurized from the lower end by nitrogen gas to fabricate the honeycomb structure (Fig. 3(b)).
Fig. 3 Fabrication process of stack-and-draw technique: (a)preparation of the preform, and (b)drawing process for the bundled fiber.
Figure 4 shows a cutting section of a fabricated fiber bundle. The bundle was composed of 245 fibers, and the total diameter was 1.06 mm. The average inner diameter of each fiber was 60 µm, and the length was 90 cm. The average thickness of the glass walls was 0.76 µm, which satisfies the design condition shown above. Based on the microscopic picture shown in Fig. 4, the standard deviation of inner diameter of each fiber was calculated as 3.0 µm excluding the outermost layer and with this variation, fluctuation of transmission loss was estimated to be lower than 30% by calculation of transmission losses of the HE11 mode based on a ray-optic method [23]. We also measured thickness variation of glass walls by using a scanning electron microscope and found the standard deviation was around 0.04 µm. This corresponds to changes in the transmission loss of around 5%. From these results, we confirmed that the irregularity in the shape of the cores of the fabricated fiber bundle will not affect the transmission property very much.
4. Experiment
Figure 5 shows the loss spectrum of the 90-cm long, tube-leaky fiber bundle measured by using an FT-IR spectrometer. In this measurement, infrared light was coupled to around 120 fibers at the center of the bundle via a metal-coated hollow optical fiber with an inner diameter of 0.7 mm. As designed, a low loss region was observed at a wavelength of 3-4 µm due to the interference effect of the thin glass wall. The loss increases in longer wavelengths because of the infrared absorption of the borosilicate glass.
Fig. 5 Loss spectrum of fabricated hollow-fiber bundle with a length of 90 cm measured by using a FT-IR spectrometer. Incident light was coupled to the bundle via a hollow-optical fiber with an inner diameter of 0.7 mm.
Figure 6 shows the measured loss spectrum of the fabricated bundle compared to that of the bundle made of around 120 pieces of silver-coated hollow optical fibers with a length of 3 cm and an inner diameter of 65 μm. It is known that the transmission loss of the hollow optical fiber is inversely proportional to the third power of the fiber’s inner diameter [18]. Therefore, the transmission losses of small diameter fibers are usually very high as shown in the spectrum of silver hollow-optical fiber in Fig. 6 even when the length is as small as 3 cm. However, the losses of bundled anti-resonant hollow fibers are much lower than those of the conventional hollow fiber bundle as shown in the figure. This is mainly because in the bundled tubular fiber, the portion of light leaking out from the core is confined to the adjacent cores and propagates again in the adjacent cores. As a result of this phenomenon, the bundled small diameter fibers seems to act as a large core hollow optical fiber in terms of power delivery medium. However, this also means that crosstalk among the fibers is reasonably high, which may cause degradation of the transmitted images.
Fig. 6 Measured loss spectrum of fabricated hollow-fiber bundle (90-cm long) compared to bundle of silver-coated hollow optical fibers with length of 3 cm. Both fiber bundles are composed of around 120 fibers with an inner diameter of 60-65 μm.
We performed a thermal image delivery experiment using the fabricated bundled fiber. In the experiment, a 0.2-mm diameter, nichrome wire heated to 120°C was used as a sample target. We placed a ZnSe lens with a focal length of 50 mm and a diameter of 25 mm at the input end of the fiber to form reduced images of the target on the input end of the bundle. The distances between the target and the lens, and the lens and the fiber’s input end were set to obtain one-third magnification power. Figure 7 shows the observed near field images of the output end of the bundle with different fiber lengths observed by an InSb infrared camera (FLIR SC6000MWIR) with a detection wavelength range of 3-5 µm, which coincides with the low-loss region of the fibers. The thickness of the nichrome wire was 200 μm, which nearly coincided with the size of each fiber considering the one-third magnification power of the lens system. We confirmed that the bundle successfully delivered thermal images although the resolution was somewhat degraded compared to the original target image.
Fig. 7 Observed thermal images of heated metal wire of 0.2-mm thickness transmitted through fiber bundles with different lengths.
To estimate degradation of image resolution during transmittance in the bundle fiber, we performed a simple numerical simulation based on ray tracing calculation. In the ray-optic model, we calculated mutual coupling of power transfer between a core and surrounding six cores numbered as shown in Fig. 8. For the ray with incident angle of θ, the power attenuation α in a core with diameter of 2T is expressed by power reflectance R at the wall considering multireflection on the inner and outer surfaces as [23],
When absorption in the glass wall is negligibly small, the power of ray that is lost from the core is distributed to the surrounding cores. Then, after transmission of distance x, the optical power in the core P’0 is expressed from the powers in the cores #0-6 before the transmission P0-P6 as,
This calculation was repeated with 0.1-mm step considering 7 surrounding layers and the power leaked outside from the outermost layer was dumped. In the simulation, we assumed an input beam with a divergence angle of 16.7° that corresponded to the launching angle to the fiber from the ZnSe lens at the input end.Figure 9 shows the change in the thickness of the observed images for different bundle lengths. The observed images were scanned in a direction perpendicular to the wire image at 5 different points, and the average thickness was calculated based on Gaussian fitting. A simulated result is also shown in the figure for comparison. In this simulation, an initial beam distribution of a linear shape was launched in to the central cores of the fiber. For both of the experiment and simulation results, the resolution rapidly degrades at small bundle lengths and it becomes constant at lengths larger than 10 cm. This is because lossy high order modes having a large propagation angle are excited in the fibers and most of them are lost nearby the input end. Then only low order modes with small leaky losses survive and deliver the image. The image resolution of the bundle with enough lengths was around 420 μm, which corresponds to almost the size of two fiber cores considering the one-third magnification of the lens system used in the experiment. The discrepancy between the experiment and the simulation seems to be due to coherent mode coupling between the cores that was not considered in the simulation.
Fig. 9 Thickness of observed thermal images of heated metal wire of 0.2-mm thickness transmitted through fiber bundles with different lengths. Results of numerical simulation based on a ray-optic method is also shown for comparison.
Although the total diameter of the fabricated fiber bundles was 1.06 mm, which was small enough to be inserted into a working channel of endoscopes, the smallest bending radius was around 40 cm due to the honeycomb structure comprised of glass. However, the fiber bundle is good for rigid endoscopes used in laparoscopic surgery. Those rigid endoscopes typically have a length of 30-50 cm and diameter of 5-10 mm with a working channel of 3-5 mm in diameter, and the proposed fiber bundle is within these dimensions. Then, we built an imaging system that is suitable for applications with rigid endoscopes. An outer appearance and the structure of the input end of the system are shown in Fig. 10. In this system, we attached a sapphire half-ball lens of 2.0-mm diameter to the end of the 30-cm long fiber bundle in place of the large ZnSe lens used in the previous experiment to reduce the outer diameter of the system. The half-ball lens was held by a stainless-steel sleeve with a diameter of 2 mm as shown in Fig. 10. The effective focal length of the lens is 1.3 mm and the lens was located between the target and the fiber bundle to obtain one-third magnification power.
Fig. 10 Outer appearance and structure of fabricated fiber bundle with half-ball lens at distal end.
We firstly checked the sensitivity of the system using a ceramic plate heated to nearly body temperature. Figure 11(a) shows an observed image of the plate at a temperature of 32.0°C that was placed in the air. We confirmed that the system had a signal-to-noise ratio that was good enough to obtain a thermal image with a temperature lower than body temperature. The image resolution was 0.3 mm for this case. Figure 11(b) shows a thermal image of two plates with temperatures of 38.0°C and 37.3°C. These two plates were placed side-by-side and were clearly differentiated, and this showed that the temperature resolution of the system was around 0.7°C in the body temperature range. This is good enough for detection of inflammation of mucosa membrane observed in early cancer tissues whose temperature elevation is usually larger than 1.0°C. The image resolution was limited to 0.45 mm in this case because of the low contrast between the two plates.
Fig. 11 Observed thermal images of heated ceramic plates: (a) plate at 32.0°C placed in air and (b) two plates with temperatures of 38.0 and 37.3°C.
Figure 12 shows an image of a looped metal wire at a temperature of 38.0°C, which was observed by the fiber bundle. The size of the sample is also shown in the figure. We acquired a clear image of the sample’s shape and the imaging resolution was 0.4 mm. We fit a Gaussian distribution to the observed image of the wire and obtained the resolution from the full-width-half-maximum of the Gaussian. From this result, we believe that the feasibility of the imaging system based on the bundle of anti-resonant hollow optical fibers for endoscopic applications for early cancer detection and laser hyperthermia is successfully shown.
Fig. 12 Observed thermal image of a looped metal wire with a diameter of 0.2 mm heated to 38.0°C (right) and the appearance of the wire (left)..
5. Conclusion
A bundle of anti-resonant hollow optical fibers was proposed for endoscopic infrared-thermal imaging. In these hollow optical fibers, light is confined to the air core of each fiber due to the anti-resonant effect of the glass wall and is transmitted as a leaky mode with relatively low loss by setting the thickness of the glass wall to a quarter-wavelength optical thickness. A bundle of 245 fibers made of borosilicate glass was fabricated by using a glass drawing technique, and the length and total diameter of the bundle were 90 cm and 1.06 mm, respectively. As designed, the bundle fiber showed low losses in the wavelength range of 3 to 4 μm that fit the detection wavelengths of InSb infrared cameras. In the thermal image delivery experiment, an image resolution of around 420 μm that corresponds to twice the size of each fiber was obtained although crosstalk between adjacent fibers was observed. By considering applications with rigid endoscopes, an imaging system composed of a 30-cm long fiber bundle and a half-ball lens with a diameter of 2 mm that can be inserted into a working channel of endoscopes was fabricated. By using an imaging system with a one-third reduction ratio, an image resolution of around 420 μm was successfully obtained with a field-of-view of 2-mm diameter. Another tests showed that the minimum detected temperature was 32.0°C, and the temperature resolution of the system was around 0.7°C.
References and links
1. M. Diakides, J. D. Bronzino, and D. R. Peterson, Medical Infrared Imaging: Principles and Practices (CRC, 2013).
2. M. Vollmer and K. Mollmann, Infrared Thermal Imaging: Fundamentals, Research and Applications (Wiley, 2010).
3. N. Arora, D. Martins, D. Ruggerio, E. Tousimis, A. J. Swistel, M. P. Osborne, and R. M. Simmons, “Effectiveness of a noninvasive digital infrared thermal imaging system in the detection of breast cancer,” Am. J. Surg. 196(4), 523–526 (2008). [CrossRef] [PubMed]
4. T. M. Buzug, S. Schumann, L. Pfaffmann, U. Reinhold, and J. Ruhlmann, “Functional Infrared Imaging for Skin-Cancer Screening,” in Conference Proceedings of IEEE Engineering in Medicine and Biology Society, (IEEE, 2006), pp. 2766–2769.
5. Q. Wang, L. Xie, Z. He, D. Di, and J. Liu, “Biodegradable magnesium nanoparticle-enhanced laser hyperthermia therapy,” Int. J. Nanomedicine 7, 4715–4725 (2012). [PubMed]
6. Y. Matsuura and K. Naito, “Flexible hollow optical fiber bundle for infrared thermal imaging,” Biomed. Opt. Express 2(1), 65–70 (2011). [CrossRef] [PubMed]
7. 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). [CrossRef] [PubMed]
8. M. A. Duguay, Y. Kokubun, T. L. Koch, and L. Pfeiffer, “Antiresonant reflecting optical waveguides in SiO2-Si multilayer structures,” Appl. Phys. Lett. 49(1), 13–15 (1986). [CrossRef]
9. N. M. Litchinitser, A. K. Abeeluck, C. Headley, and B. J. Eggleton, “Antiresonant reflecting photonic crystal optical waveguides,” Opt. Lett. 27(18), 1592–1594 (2002). [CrossRef] [PubMed]
10. W. Belardi and J. C. Knight, “Hollow antiresonant fibers with low bending loss,” Opt. Express 22(8), 10091–10096 (2014). [CrossRef] [PubMed]
11. J. R. Hayes, F. Poletti, M. S. Abokhamis, N. V. Wheeler, N. K. Baddela, and D. J. Richardson, “Anti-resonant hexagram hollow core fibers,” Opt. Express 23(2), 1289–1299 (2015). [CrossRef] [PubMed]
12. A. N. Kolyadin, A. F. Kosolapov, A. D. Pryamikov, A. S. Biriukov, V. G. Plotnichenko, and E. M. Dianov, “Light transmission in negative curvature hollow core fiber in extremely high material loss region,” Opt. Express 21(8), 9514–9519 (2013). [CrossRef] [PubMed]
13. S. Kobayashi, T. Katagiri, and Y. Matsuura, “Microstructured tube-leaky glass waveguide for delivery of high-powered Er:YAG laser,” J. Lightwave Technol. 32(5), 986–990 (2014). [CrossRef]
14. A. Urich, R. R. J. Maier, F. Yu, J. C. Knight, D. P. Hand, and J. D. Shephard, “Silica hollow core microstructured fibres for mid-infrared surgical applications,” J. Non-Cryst. Solids 377, 236–239 (2013). [CrossRef]
15. S. Février, B. Beaudou, and P. Viale, “Understanding origin of loss in large pitch hollow-core photonic crystal fibers and their design simplification,” Opt. Express 18(5), 5142–5150 (2010). [CrossRef] [PubMed]
16. L. Vincetti and V. Setti, “Waveguiding mechanism in tube lattice fibers,” Opt. Express 18(22), 23133–23146 (2010). [CrossRef] [PubMed]
17. M. Miyagi and S. Nishida, “A proposal of low loss leaky waveguide for submillimeter waves transmission,” IEEE Trans. Microw. Theory Tech. 28(4), 398–401 (1980). [CrossRef]
18. M. Miyagi, “Bending losses in hollow and dielectric tube leaky waveguides,” Appl. Opt. 20(7), 1221–1229 (1981). [CrossRef] [PubMed]
19. Y. Matsuura, R. Kasahara, T. Katagiri, and M. Miyagi, “Hollow infrared fibers fabricated by glass-drawing technique,” Opt. Express 10(12), 488–492 (2002). [CrossRef] [PubMed]
20. R. Kasahara, Y. Matsuura, T. Katagiri, and M. Miyagi, “Transmission properties of infrared hollow fibers produced by drawing a glass-tube preform,” Opt. Eng. 46(2), 025001 (2007). [CrossRef]
21. E. D. Palik, ed., Handbook of Optical Constants of Solids (Academic, 1985)
22. A. A. Rifat, G. A. Mahdiraji, D. M. Chow, Y. G. Shee, R. Ahmed, and F. R. M. Adikan, “Photonic crystal fiber-based surface plasmon resonance sensor with selective analyte channels and graphene-silver deposited core,” Sensors (Basel) 15(5), 11499–11510 (2015). [CrossRef] [PubMed]
23. Y. Matsuura, A. Hongo, M. Saito, and M. Miyagi, “Loss characteristics of circular hollow waveguides for incoherent infrared light,” J. Opt. Soc. Am. A 6(3), 423–427 (1989). [CrossRef]
