IEEE Proof. Low-Cost Transducer Based On Surface Scattering Using Side-Polished D-Shaped Optical Fibers. Volume 7, Number 5, October 2015

Transcription

1 Low-Cost Transducer Based On Surface Scattering Using Side-Polished D-Shaped Optical Fibers Volume 7, Number 5, October 2015 Y. S. Ong W. Kam S. W. Harun R. Zakaria Waleed S. Mohammed DOI: /JPHOT Ó 2015

2 Photonics Journal Low-Cost Transducer Based On Surface Scattering Using Side-Polished D-Shaped Optical Fibers Y. S. Ong, 1,2,3 W. Kam, 1,2,3 S. W. Harun, 2 R. Zakaria, 2 and Waleed S. Mohammed Department of Physics, University of Malaya, Kuala Lumpur 50603, Malaysia 2 Photonic Research Centre, Faculty of Science, University of Malaya, Kuala Lumpur 50603, Malaysia 3 Center of Research in Optoelectronics, Communications and Control Systems (BU-CROCCS), Bangkok University, Bangkok 10120, Thailand DOI: /JPHOT Ó Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires permission. See for more information. Manuscript received August 24, 2015; revised September 17, 2015; accepted September 17, Date of publication Month 00, 0000; date of current version Month 00, This work was supported by the Research University Grant RU and Grant UM.C/625/1/HIR/MOHE/SCI/01. Corresponding author: R. Zakaria ( rozalina@um.edu.my). Abstract: This work presents a new low-cost and environmentally friendly optical transducer based on surface scattering in a side-polished fiber mounted on a glass groove. The surface roughness caused by polishing is used to estimate the change in refractive index on top of the fiber. Changing the surrounding index affects the scattering properties and, consequently, the leakage of the guided mode. This effect is experimentally investigated through the change in the attenuation coefficient when the surrounding refractive index is changed by dropping different weight ratios of glucose solutions onto the polished side. The measured change of the attenuation coefficient is consistent with the finite-element calculations. The effect of different polished depths was theoretically investigated to optimize the working conditions. Index Terms: Author, please supply index terms/keywords for your paper. To download the Taxonomy go to 1. Introduction Optical-fiber transducers are commonly used as sensors in various sensing fields such as chemical sensing [1] and humidity sensing [2], the automotive industry, and even molecular biotechnology analysis monitoring. Studies on this technology have drawn much attention to produce high-durability sensors that can be used in hazardous environment and extreme conditions such as extreme temperature or corrosive medium. Numerousstructuresapplythisoptical-fiber transduction, such as Bragg grating in optical fibers, tapered fiber optics [2], D-shaped fiber optics [3], [4], reflection-based sensors, and ring- or knot-based fiber optics [5]. An optical fiber operates as a transducer when a fraction of power and the evanescent field propagate in the outer environment. Hence, the sensing part of the fiber is sensitive to the variation of the surrounding medium, which can be detected through the loss of light energy or phase changes. Tapered fibers provide a low confinement of light because of the small radius at the waist, which is the main reason that it is sensitive to any change in the outer environment. Tapered fiber has been implemented in a large number of applications such as optical filtering [6], coupling [7], AQ Vol. 7, No. 5, October 2015 Page 1

3 Photonics Journal Fig. 1. Scattering effect of reflection at the rough interface for different refractive indices. The intensity of the reflected light and scattered light is represented by the thickness of the arrow (not to scale). The dotted line represents the remaining cladding or rough core surfaces. and refractive index sensing [8], [9] because of its flexibility. Nonetheless, D-shaped fiber is an asymmetrical structure, which is formed by polishing one side of the fiber to the core. This asymmetry enables a few applications of the fiber, such as in intermodal interferometers [10] and polarization-maintaining fibers [11]. The reduction of light confinement in the D-shaped fiber increases the evanescent field in the surrounding, which confirms that the D-shaped fiber is a potential fiber for monitoring the change in refractive index [12]. In practice, some applications require sensors that can be exposed to unprotected environments. D-shaped fibers are typically carefree transducers that require no packaging [13] for protection from dusty environments [14] and do not undergo undesired dynamic bending [15] because they are easy to clean and are rigidly supported by a substrate, respectively. This paper introduces a low-cost fiber transducer based on the change of the light-scattering properties on the surface of side-polished single-mode fiber in changing environments. Increasing the cladding's refractive index reduces the index difference between the scatterers (which are formed because of the surface roughness) and the surrounding. Thus, the scattering coefficient decreases, and there is less light leakage during propagation in the fiber core, as illustrated in Fig. 1. The sensitivity of this change depends on the polishing depth and surface roughness condition of the polished fiber. It is worth mentioning that the proposed D-shape fiber device shows high mechanical stability. Manipulating the scattering effects makes the fabrication process of this device as cost-effective as the no-coating or vacuum systems. 2. Experiment Methodology 2.1. Device Fabrication The process begins with the groove fabrication. The following is a simple scheme for making the fiber groove. Two pieces of tape are placed on a glass slide under a microscope to allow for a small gap between the pieces of tape, as shown in Fig. 2(a). The desired gap is approximately 125 m, in order to fit the optical fiber. This stage may require several repetitions because of the manual technique of fiber placement. During the process, both sides of the gap are examined under the microscope to ensure a parallel groove. The gap is subsequently filled with a 10 : 1 volume ratio of hydrofluoric acid to hydrochloric acid to etch the glass substrate. With proper timing, this step creates a smooth groove with a depth that fits half of the optical fiber [16], as shown in Fig. 2(b) and (c). Rough groove is not desired as it may produce obvious grain that could compress or bend the optical fiber. The groove was etched for 5 minutes to achieve a Vol. 7, No. 5, October 2015 Page 2

4 Photonics Journal Fig. 2. Fabrication process of a cheap transducer device. (a) Taping the groove. (b) Mixture of HCL and HF solution. (c) Etching process. (d) Epoxy fiber on the groove. (e) Polishing characterization setup. (f) Fiber subjected to in-house polishing tool. 70 m depth. The entire etching process was performed in a sonicator to ensure consistent etching along the groove. A cleaved piece of a single-mode optical fiber (SMF) SMF-28 was subsequently mounted in the groove using epoxy. The mounted fiber was left in open air to cure. A glass substrate was used because of its chemically inert properties and mechanical strength compared to its polymer counterpart [17]. Moreover, it can be improvised into microfluidic or lab-on-chip platforms because of its biocompatibility and optical transparency [18]. U-shaped grooves were created in this process because of the gap's width between the scotch tape and the isotropic etching on glass using the acid mixture solution. After curing, the fiber that was mounted in the groove was subsequently polished using 1200-grit polishing fine sandpaper. To provide some consistency in the polishing process, an inhouse polishing tool was made by modifying a USB fan, where the blades were removed and the sand-polishing sheet was fixed around the center. The polishing process ended after the desired fiber diameter was roughly determined. The rough determination of desired fiber diameter was done by manual observation of red light leakage during polishing process. Red light is feed into the optical fiber during the process and the process is terminated once the red light is observed from polished surface. The fiber was observed using an optical microscope afterward to measure the diameter of the polished area. This process provides an indication of the distance between the core centers and the interface. Note that the distance between the core and the interface must be less than 6 m to ensure the presence of evanescent wave at the sensing region. The depth of polishing can be controlled using the fabricated groove. Depth of the groove will influence the distance of polishing which the process will stop when the sand paper touch the groove. To better estimate the actual distance to the core from the surface, the loss in fiber was measured and compared to the simulations as described later. The estimated length of the polished fiber is approximately 15 mm, which acts as the sensing part of the device. Sandpapers of different coarseness grades were used to assist in the polishing process, and the roughness was measured using a Dektak 150 Surface Profiler. The entire fabrication process is shown in Fig Experiment Setup Fig. 3 shows the optical setup used to characterize the fiber optic transducer. The amplified spontaneous emission (ASE nm) was used as an input source, and the output was coupled to an optical spectrum analyzer (OSA) to observe the changes. Deionized (DI) water and different weight ratio of glucose solutions with different refractive indices of were placed on the polished surface of the optical fiber. The transmittance was recorded to analyze the effect of the changes on the measured signal Vol. 7, No. 5, October 2015 Page 3

5 Photonics Journal Fig. 3. Experimental setup for fiber optic sensor characterization. Red light confirms the sensitive area. The lower right microscope photo measures the diameter of the polished area. Solutions of different refractive indices were controlled by mixing different concentrations of distilled water and glucose. The refractive index of the solutions for two wavelengths was measured using the Prism Coupler and Loss Measurement SPA-4000 [19], as shown in Table Results and Discussion 3.1. Experimental TABLE 1 Refractive index of distilled water and glucose solutions at nm and 1550 nm wavelengths In this experiment, changes of the refractive index affect the scattering properties of the polished side of the fiber. This effect was measured based on the change in the attenuation coefficient of the light while propagating in the fiber. Extracting the exact coefficient for each index requires the breaking of the sample at different lengths (cut-back method). Another approach is to use the scattering effects for side coupling [20]. The relative scattering attenuation coefficient ðþ was measured by covering the polished side at different lengths. The coverage length ðl so Þ was also measured from microscope images that were obtained using an optical microscope, which was placed on top of the fiber. The output power for each length was also recorded. From the output intensity, the attenuation coefficient was subsequently calculated using Beer Lambert Law I ¼ I o e L : (1) Considering that different refractive indices were dropped onto the D-shape fiber, the relation of the coefficient can be written as ln I so ¼ð air so Þl so air ðl air Þþ2 tp l tp : (2) I o Vol. 7, No. 5, October 2015 Page 4

6 Photonics Journal Fig. 4. Diagram of a side-polished D-shape fiber with the polished length l air and tapered length l tp. AQ2 Fig. 5. Graph of ln ði j =I o Þ versus the amount of DI water on the polished part of the D-shaped fiber (Sample A). The first term of (2) on the right-hand side represents the attenuation difference from the solution and air on a sensing part with length l so. air is the attenuation coefficient of the polished part of the fiber ðl air Þ due to air, and so is the attenuation coefficient of the D-shaped fiber that was covered by the solution. tp is the attenuation coefficient of the tapered part ðl tp Þ.Thefirst term should yield a positive value if the scattering attenuation reduces. I 0 is the reference intensity of the ASE signal in air. I so is the intensity of the transmitted light when a liquid with a certain refractive index and length of coverage l so is also placed on the sensing part of the fiber. In this experiment, few samples were prepared using the proposed sand-polishing method. The samples were labeled from A to D. Fig. 5 shows a linear-scale measurement change with the transmittance (signal/reference or I so =I 0 ) against the coverage length for sample A. The relative scattering attenuation coefficient ð air so ) was extracted using the analogy of the linear fit of the measurements in Fig. 5 to (2). This approach was subsequently used for weight ratio of glucose solution, and the relative attenuation coefficient for each retrieved sample is shown in Fig. 6. Fig. 6 shows that the relative attenuation coefficient increases with the refractive index of the solution because of the decrease in core confinement when the index of the upper region increases. Stronger evanescent waves appear in the sensing region; hence, the scattering effect increases. Note that different weight ratio of glucose solution affect the solution index. The sensitivity of the measurements in terms of the relative attenuation coefficient per refractive index unit (RIU) can be estimated from the slope of the linear fit. For the data in Fig. 6, the estimated sensitivity is approximately 1130 m 1 per refractive index unit. The large difference among Vol. 7, No. 5, October 2015 Page 5

7 Photonics Journal Fig. 6. Relative attenuation coefficient ð air sample Þ versus the refractive index of the solution from Sample A. Fig. 7. Core dimensions and claddings for COMSOL stimulation sweeping. (a) Different polishing depth. (b) Placing different solutions on the D-shaped fiber. these samples is attributed to the inaccuracy of the manual approach in determining the polishing depth and the imaging resolution, as previously mentioned. To investigate this effect, the following subsection finite-element method (FEM) was used to calculate the polishing depth effect on the modal power confinement [22] Numerical Analysis Here, different polishing conditions were considered, as shown in Fig. 7(a). Two main parameters were investigated: the polishing depth and the surrounding refractive index. Using FEM, the field profiles and modal effective index for each case were calculated, as shown in Figs. 8 and 9. The calculated effective indices were subsequently used to extract the attenuation coefficient caused by the surface roughness as with surface scattering [22]: ¼ 42 3 (3) h þ 2 where is the root mean square surface roughness, is the transverse propagation constant in qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi the core region ð ¼ k0 2n2 2 Þ, is the transverse propagation in the cladding region qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð ¼ 2 k0 2n2 Þ, is the mode propagation constant, and h is the height of the remaining core. Surface roughness, mode's propagation constant and core dimension are factors that dictate surface scattering strength in this attenuation coefficient relationship. Rougher surface able to enhance surface scattering phenomenon, as it is stronger scatterer. The relative Vol. 7, No. 5, October 2015 Page 6

8 Photonics Journal Fig. 8. Mode profiles for different polishing depths with air on top (the white dashed line represents the outline of the core and the polishing plane). Fig. 9. Propagation constant versus the core dimension for different claddings. attenuation coefficient was found by subtracting the attenuation coefficient when air was the surrounding from that of the solution (water and different weight ratio of glucose solutions, as shown in Fig. 7(b)). In the calculations in Fig. 8, the 1531 nm wavelength was selected. The polishing was assumed to be uniform, and the fiber was considered to be straight. The refractive indices of the core and cladding were and , respectively. The different media with different refractive indices are listed in Table 1. This value shows that the polishing plane was closer to the core as it approached zero (when the polishing plane reached the fiber's center.) Light confinement inside the core of the propagating mode from the electric field distribution is shown in Fig. 8. As shown in the figures, light tends to spread in the substrate region when the polishing depth increases. Polishing the fiber to the core region enables a weak evanescent wave to Vol. 7, No. 5, October 2015 Page 7

9 Photonics Journal Fig. 10. Attenuation coefficient of the rms roughness nm (Sample B) based on the calculation of the effective index by COMSOL. TABLE 2 Roughness, diameter, polishing depth, and sensitivity of the samples appear in the sensing region. In addition to the reduction of power confinement, the fundamental mode splits into two with different propagation constant because of the induced asymmetry to the fiber (see Fig. 9). The difference of two propagation constants becomes observable when the fiber becomes more asymmetric, which is favorable for polarization maintenance. The attenuation coefficient because of surface scattering was calculated using (3) and plotted in Fig. 10 for two modes of polarization (HE x and HE y ) versus the distance from the core center at two different surrounding refractive indices. The graph in Fig. 10 shows a non-linear increment of the relative attenuation coefficient with the polishing depth. The relative attenuation appears to increase more rapidly after a distance of 4 m when the core diameter decreases. This distance is approximately the radius of the core, which shows that the core begins to be exposed, which increases the effect of the surrounding on the guiding-mode properties and the scattering properties of the surface roughness. When a higher-refractive-index solution is placed on the interface, the evanescent wave increases in the sensing region because of the reduction of the attenuation coefficient, and the guided light is pushed up towards the surface. This phenomenon is consistent with the positive slope in Fig. 6. As previously mentioned, three samples were prepared with similar roughness and different polishing depth values, as shown in Table 2. Here, the change in performance is obviously because of the effect of the polishing depth. The surface roughness values in Table 2 ðr q Þ were measured using a surface profiler and the estimated distance core center from the interface come from measured diameter of polished area. The estimated distance core center from the interface was compared with the simulation result to reconfirm the polishing depth. The measured etched depth, which was obtained from experimental data, is consistent with the simulation using (3), as shown in Fig. 11 for sample A. This comparison between using HE y polarization and using both polarizations is not significant. Samples B and C have larger distances from the corecenter.hence,theevanescentwave strength is expected to decrease, as clearly shown by the reduction of their slopes in Fig Vol. 7, No. 5, October 2015 Page 8

10 Photonics Journal The graphs were compared with the calculations based on the surface scattering and effective indices, which were extracted using the FEM method. The graph is consistent with the experimental measurements. The trend of all samples increases with the refractive index changes. However, the sample nearest the core center exhibits the highest sensitivity. The results are tabulated in Table Conclusion This cost-efficient and easy-fabrication optical-fiber transducer provides a valuable method to detect changes of the refractive index using the relative attenuation from surface scattering. The polishing depth plays an important role in enhancing the sensitivity of the transducer. This aspect has been theoretically discussed and well demonstrated. In the future, it is sensible to fabricate and investigate different techniques to change surface roughness and its effect on the output result. References Fig. 11. Fitting for Sample A at 150 g/l concentration (HEy). Fig. 12. Relative scattering attenuation from the stimulation (HEy) for the estimated depth and roughness for the corresponding sample and experiment data. [1] X.-D. Wang and O. S. Wolfbeis, Fiber-optic chemical sensors and biosensors ( ), Anal. Chem., vol. 85, no. 2, pp , Jan [2] J. M. Corres, F. J. Arregui, and I. R. Matias, Design of humidity sensors based on tapered optical fibers, J. Lightw. Technol., vol. 24, no. 11, pp , Nov Vol. 7, No. 5, October 2015 Page 9

11 Photonics Journal [3] M.-H. Chiu, C.-H. Shih, and M.-H. Chi, Optimum sensitivity of single-mode D-type optical fiber sensor in the intensity measurement, Sens. Actuators B, Chem., vol. 123, no. 2, pp , May [4] C.-H. Chen, T.-C. Tsao, J.-L. Tang, and W.-T. Wu, A multi-d-shaped optical fiber for refractive index sensing, Sensors, vol. 10, no. 5, pp , May [5] K. S. Lim et al., Current sensor based on microfiber knot resonator, Sensors Actuators A, Phys., vol. 167, no. 1, pp , May [6] X. Jiang, Y. Chen, G. Vienne, and L. Tong, All-fiber add-drop filters based on microfiber knot resonators, Opt. Lett., vol. 32, no. 12, pp , [7] A. Boucouvalas and G. Georgiou, External refractive-index response of tapered coaxial couplers, Opt. Lett., vol. 11, no. 4, pp , [8] C. Bariáin, I. R. Matías, F. J. Arregui, and M. López-Amo, Optical fiber humidity sensor based on a tapered fiber coated with agarose gel, Sensors Actuators B, Chem., vol. 69, no. 1/2, pp , Sep [9] M. Zibaii et al., Non-adiabatic tapered optical fiber sensor for measuring the interaction between -amino acids in aqueous carbohydrate solution, Meas. Sci. Technol., vol. 21, no. 10, Sep. 2010, Art. ID [10] L. Gao, T. Zhu, J. Zeng, and K. S. Chiang, Temporal response measurement of magnetic fluids based on D-shaped fiber intermodal interferometer, Appl. Phys. Exp., vol. 6, 2013, Art. ID [11] C. Ruiz Zamarreño, P. Zubiate, M. Sagües, I. Matias, and F. Arregui, Experimental demonstration of lossy mode resonance generation for transverse-magnetic and transverse-electric polarizations, Opt. Lett., vol. 38, no. 14, pp , Jul [12] A. O. Dikovska et al., Optical sensing of ammonia using ZnO nanostructure grown on a side-polished optical-fiber, Sensors Actuators B, Chem., vol. 146, no. 1, pp , Apr [13] K. Lim, S. Harun, H. Arof, and H. Ahmad, Fabrication and applications of microfiber, in Selected Topics on Optical Fiber Technology, D. M. Yasin, Ed. Rijeka, Croatia: InTech, 2012, pp [14] M. Fujiwara, K. Toubaru, and S. Takeuchi, Optical transmittance degradation in tapered fibers, Opt. Exp., vol. 19, no. 9, pp , Apr [15] I. Matias, Transmitted optical power through a tapered single-mode fiber under dynamic bending effects, Fiber Integr. Opt., vol. 22, no. 3, pp , [16] C. Iliescu, K. L. Tan, F. E. Tay, and J. Miao, Deep wet and dry etching of Pyrex glass: A review, Proc. ICMAT, 2005, pp [17] C. Iliescu, H. Taylor, M. Avram, J. Miao, and S. Franssila, A practical guide for the fabrication of microfluidic devices using glass and silicon, Biomicrofluidics, vol. 6, no. 1, Mar. 2012, Art. ID [18] J. M. Nagarah and D. A. Wagenaar, Ultradeep fused silica glass etching with an HF-resistant photosensitive resist for optical imaging applications, J. Micromech. Microeng., vol. 22, no. 3, Mar. 2012, Art. ID [19] W. Lim et al., Graphene oxide-based waveguide polariser: From thin film to quasi-bulk, Opt. Exp., vol. 22, no. 9, pp , May [20] H. Fallah et al., Demonstration of side coupling to cladding modes through zinc oxide nanorods grown on multimode optical fiber, Opt. Lett., vol. 38, no. 18, pp , Sep [21] J. A. Tamada, M. Lesho, and M. J. Tierney, Keeping watch on glucose, Spectrum, vol. 39, no. 4, pp , Apr [22] C. Viphavakit et al., Realization of a polymer nanowire optical transducer by using the nanoimprint technique, Appl. Opt., vol. 53, no. 31, pp , Nov Vol. 7, No. 5, October 2015 Page 10