Humidity sensing using plastic optical fibers

HUMIDITY SENSING USING PLASTIC OPTICAL FIBERS C. M. Tay, K. M. Tan, S. C. Tjin, C. C. Chan, and H. Rahardjo School of Electrical and Electronic Engineering Nanyang Technological University 50 Nanyang Avenue Singapore 639798, Singapore Received 11 May 2004 ABSTRACT: The use of plastic optical fibers (POFs), curved at the sensing point and coated with cobalt chloride (CoCl2) and gelatin as the overlay material, for relative humidity sensing is demonstrated in this paper. The fiber-core diameter and bending radius of the sensing point affects the sensitivity of the sensor to a great extent. © 2004 Wiley Periodicals, Inc. Microwave Opt Technol Lett 43: 387–390, 2004; Published online in Wiley InterScience (www.interscience.wiley. com). DOI 10.1002/mop.20479 Key words: plastic optical fiber; fiber-optic sensor; humidity sensing 1. INTRODUCTION Humidity sensing is important in various areas such as industrialprocess control, buildings for human comfort, and geotechnical measurements. In industries as well as humidity-controlled areas where humans reside or work, in which product cost and equipment complexity need to be kept low, a simple and low-cost humidity sensor is very desirable. In geotechnics, with the help of simple and effective instrumentation that can measure humidity in soils, the total suction of soils can be determined in the laboratory or in the field and can be used to compute the shear strength of soil. Hence, development of a relative humidity (RH) instrument is an important field of study and many kinds of fiber-optic relativehumidity sensors have been proposed [1–9]. Fiber-optic sensors have various advantages over conventional methods of humidity measurement, such as electrical-resistance, capacitance, thermal-conductivity, and wet– dry-bulb methods [10]. The advantages include immunity to electromagnetic interference (EMI), small size, noncorrosiveness, the ability to perform long-term sensing over great distances, and so forth. Fiber-optic relative-humidity sensors that have been reported include the use of a porous thin-film interferometer deposited on the tip of a fiber [1]. The absorbed water changes the refractive index of the thin films and thus the reflectivity of the interferometer. The resultant modulation of the reflected optical intensity is detected. The measurable humidity range is 0%RH to 80%RH. Another fiber-optic humidity sensor reported is based on the deposition of a hydrophilic gel (agarose) on a biconically tapered optical fiber [2]. The optical power transmitted through the taper varies as a function of the humidity because of the humidity-induced refractive-index change in the gel surrounding the tapered zone. The effective sensing range is 30%RH to 80%RH. Another example of a fiber-optic humidity sensor was presented by Kronenberg et al. [3], where polyimide recoated fiber Bragg gratings (FBGs) were used and tests showed that the sensor has a linear, reversible, and accurate response from 10%RH to 90%RH. Russell et al. [4] used cobalt chloride (CoCl2) as the colorimetric reagent to detect the presence of humidity that was held to the 600-␮m diameter plastic clad silica (PCS) fiber by gelatin, the overlay material. Light at 680 nm filtered by a monochromator was passed into the fiber and the humidity response of the sensor was reported to be from 60%RH to 80%RH. Hypszer et al. [5] used a light emitting diode (LED) at 690 nm as a light source for a straight 200-␮m PCS fiber coated with CoCl2 and polyvinyl alcohol (PVA). The humidity sensing range was 60%RH to 100%RH. Bownass et al. [6] used gelatin as the humidity-sensing film coated on the cladding of standard telecommunications single-mode fiber (SMF). The whispering gallery (WG) mode interacts with the film at the cladding-overlay interface and causes Fresnel reflection. As humidity changes, the gelatin layers swell and the refractive index decreases. The power of the back-reflected light was measured and the humidity response of the sensor starts from 62%RH to the higher humidities. Otsuki et al. [7] found that bending fiber to a small radius improved the response of a sensor drastically, as did Jindal [8]. Jindal et al. used CoCl2/PVA as the sensing film for a curved 200-␮m PCS fiber and obtained a sensing range of 3%RH to 90%RH. A He-Ne laser source at 632.8 nm was used. Muto et al. [9] used a straight plastic optical fiber (POF) coated with hydroxyethylcellulose (HEC) and polyvinylidenefluoride (PVDF) as the sensing material. The humidity sensing range of the sensor was 20%RH to 90%RH. In this paper, making use of the advantages of large-diameter plastic fiber, higher sensitivity due to fiber bending and low-cost humidity-sensitive reagents, a simple and low-cost POF sensor based on cobalt chloride (CoCl2) and gelatin coating on the curved sensing point with a humidity sensing range from 60%RH to 95%RH is presented. An investigation into the effect of bending the radius of the fiber at the sensing point as well as fiber-core diameter upon the sensitivity of the humidity sensor is conducted in order to find the best way to improve sensing performance. The background theory is given in section 2, the experimental results and discussions are given in section 3, and the conclusion is presented in section 4. 2. BACKGROUND THEORY CoCl2 interacts with water vapor in a well-known colorimetric interaction to change its colour from deep blue (anhydrous state) to pink (hydrated state). The chemical formula of hydrated cobalt chloride is CoCl2 䡠 6(H2O). The anhydrous form has high optical absorption between 550 and 750 nm, with the peak at 690 nm [4]. The hydrated form has little or no absorption. CoCl2 forms salt crystals that adhere poorly to the polymethyl methacrylate (PMMA) (n ⫽ 1.489 at ␭ ⫽ 680 nm [9], where n is the refractive index of PMMA and ␭ is the wavelength of light) core of the POF. An overlay material that is optically clear and hydrophilic is needed, so that the fiber evanescent wave can interact with the sensing film and water vapor can diffuse into the film and interact with the CoCl2 to change its absorbing properties. Gelatin is a suitable material, as CoCl2 dissolves readily in it to form a transparent film on the fiber. Dry gelatin has a refractive index of 1.5, found by ellipsometry [4], but when gelatin absorbs moisture it swells and its density decreases, leading to a reduction in refractive index of the film [6]. When light from the source travels through the fiber, depending on the humidity and thus the spectral absorption of the CoCl2, the attenuated light intensity will be [5]: I ⫽ I 0exp ⫺␣rLc , D (1) where I 0 is the input-light intensity, ␣ is the absorption coefficient of CoCl2, r is the thickness of reagent layer, L is the length of sensing segment, c is the reagent concentration, and D is the diameter of the uncladded optical fiber. MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 43, No. 5, December 5 2004 387 Figure 1 Experimental setup for relative-humidity sensing using POF 3. EXPERIMENTAL RESULTS AND DISCUSSION The experimental setup used is shown in Figure 1. An aluminum gallium arsenide (AlGaAs) light-emitting diode (LED) from Kingbright was used as the light source. It has a peak wavelength of 660 nm, a spectral-line halfwidth of 20 nm, and a viewing angle of 20°. The POF used has a core of PMMA and a very thin cladding of fluorinated polymer. Due to the large 2-mm fiber diameter, no connectors were used. The plastic fiber was bent to a radius of 1 cm at the sensing end using a flame. A momentary application of heat to the POF can sufficiently soften it to bend it around a mandrel. Powdered cobalt hexahydrate (3% by weight) and gelatin (5% by weight) was mixed with distilled water to form an aqueous solution that is heated in an oven to 65°C for 10 min. The curved fiber end is then dipped into the solution and then left out in the open to dry. The dipping was done three times to obtain three coats. The measured thickness of the film was 30 ␮m, using a vernier caliper that is accurate to 20 ␮m in order to measure the diameter of the coated fiber. A large-area (1 cm2) silicon photodetector was used to detect the optical-intensity variations from the POF and a digital multimeter (HP34401A) displays the corresponding voltage. The sensing end of the POF was placed into a climate control chamber (Feutron KPK35). The chamber’s temperature accuracy is ⫾0.1°C and the humidity resolution is 0.1%RH. The temperature within the chamber was fixed at 25°C for all the experiments and the humidity was varied from 47%RH to 95%RH to test the response of the sensor. Figure 2 shows the hysteresis response of the POF sensor. There is a response from the sensor beginning from 60%RH and terminating at 95%RH. The linear region is from 65%RH to 85%RH, the gradient of the increasing humidity curve from 65%RH to 85%RH is 0.07425 V/%RH, and the corresponding gradient for the decreasing humidity curve for the same range is 0.07485 V/%RH. The hysteresis error is only 0.0006 V, a negligible amount in terms of humidity. A change of 0.01%RH is detectable, as the digital multimeter is capable of displaying voltages up to the fourth decimal place. The sensor responds quickly to changes in humidity, in terms of seconds, and the voltage readout stabilizes to a constant value within a minute. The equilibrium time is humidity dependent. The higher the humidity is, the longer the readings take to stabilize. In order to test the repeatability of the response of the sensor, the humidity within the chamber was ramped up and back down three times. Figure 3 shows a comparison of the results of the three tests for increasing humidity and Figure 4 shows a comparison of the results of the three tests for decreasing humidity. 388 Figure 2 Humidity-sensing range and hysteresis response of the POF sensor The experimental data were fitted to a curve and the standard deviation for the increasing humidity curve was found to be about 0.05 V, an uncertainty in measured relative humidity in the linear region of about ⫾0.67%RH. The repeatability error is about 2.7% for the whole range of voltage response due to the humidity change. Likewise, the standard deviation for the decreasing humidity curve was found to be 0.02 V, an uncertainty in measured relative humidity in the linear region of about ⫾0.27%RH. The repeatability error is about 1.1% for the whole range of voltage response due to the humidity change. The POF is a lossy fiber; hence, even without bending the fiber, some light can be observed to leak from the side of a straight fiber. Bending the fiber causes an increase in bending loss at that point and also shifts the peak of the mode-field distribution away from the fiber axis towards the outer edge of the bend (Fig. 5). This causes a stronger evanescent wave to propagate through the thin film. The evanescent wave easily penetrates the thin cladding of the POF to reach the film. Therefore, a higher-intensity modulation Figure 3 Repeatability test of the humidity-sensor response for the increasing-humidity cycle MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 43, No. 5, December 5 2004 Figure 4 Repeatability test of the humidity-sensor response for the decreasing-humidity cycle can occur due to the increased interaction between the evanescent wave and the absorbing thin film that acts as a cladding for the POF. As will be demonstrated, a different bending radius will affect the sensitivity of the sensor. The humidity-sensing range is fixed by the properties of the CoCl2/gelatin film, as observed by Russell et al. [4], where there is little or no water absorbed from 30%RH to 50%RH. Apart from the higher optical-intensity modulation expected from bending the fiber, an increase in the intensity of the evanescent wave will improve the sensitivity of the sensor as well. For fibers with smaller core diameters, the evanescent tail of the core mode cannot easily penetrate the cladding, as the intensity of the core mode is smaller and the cladding of the fiber is thicker with respect to the smaller core size. A higher intensity of the evanescent wave will also allow a greater modulation with the absorption cladding film because more light can be absorbed. One way to improve the optical intensity of the evanescent wave in a POF is to use larger-diameter fibers that can carry higher optical power and also have larger core-to-cladding ratios. An experimental investigation was conducted to determine the effect of different fiber parameters, such as fiber-core diameter and bending radius, on the sensitivity of the POF humidity sensor. Figure 5 Mode-field distribution and light paths at the bent region of the POF sensing portion Figure 6 Sensor response for different fiber-core diameters Figure 6 shows the effect of different fiber diameters and Figure 7 shows the effect of different bending radius on the response of the POF sensor. It can be observed from Figure 6 that the larger the fiber diameter is, the better the sensitivity of the sensor with respect to humidity will be. All sensors used in the fiber-diameter experiment have constant bending radius of 2 mm, a constant coating thickness of 10 ␮m, and a fixed composition of 5% gelatin and 3% CoCl2 by weight. Figure 7 shows that the smaller the bending radius of the sensing portion of the fiber, the better the sensitivity with respect to humidity. In this experiment, the fibers used have a fixed diameter of 2 mm, a constant coating thickness of 10 ␮m, and a fixed composition of 5% gelatin and 3% CoCl2 by weight. There are some issues to be resolved before the sensor can be used effectively. There were some difficulties in coupling light from the LED to the fiber without connectors. In order to obtain the best coupling, the fiber should be mounted as close to the LED and photodetector as possible, otherwise there will be discrepancies in the voltage response of the sensor. There is also the problem of bending loss, as the sensor’s method of sensing is intensity modulation by the humidity-sensing film. Hence, any bending of the Figure 7 Sensor response for different bending radii MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 43, No. 5, December 5 2004 389 fiber will change the intensity of the light detected. In the experiment, the fiber lengths are kept short so that unnecessary bending is minimized, but if the sensor is to be used in the field, the probe will consist of the sensing portion of the fiber and the rest of the fiber will need to be long in order to reach the data-collection unit. Bending the fiber will lead to a change in the recorded intensity that is not due to humidity change. A possible solution to this problem may be to use multiwavelength referencing for the sensor, where the sensing film absorbs one wavelength and the second wavelength is not in the absorption range of the film. Hence, intensity changes due to fiber bending and not due to humidity changes can be eliminated. 4. CONCLUSION A POF coated with CoCl2 and gelatin as the overlay material, and interrogated using a simple LED and photodetector setup, has been demonstrated. This POF relative-humidity sensor has a sensing range from 60%RH to 95%RH with a resolution of 0.01%RH. There is negligible hysteresis error and the repeatability error can be as low as 1%. It was found that increasing the fiber diameter allows the evanescent wave to be stronger and thus increase the sensitivity of the sensor. Bending the radius affects the sensitivity of the sensor as well, with smaller radii leading to greater penetration depths of the evanescent wave in the absorption film. This leads to greater absorption of light and thus a higher opticalintensity modulation. Before the sensor can used effectively, some issues such as the coupling of light into and out of the fiber and bending loss due to handling will need to be resolved. REFERENCES 1. F. Mitschke, Fiber-optic sensor for humidity, Optics Lett 14 (1989), 967–969. 2. C. Bariain, I.R. Matias, F.J. Arregui, and M. Lopez-Amo, Optical fiber humidity sensor based on a tapered fiber coated with agarose gel, Sensors and Actuators B 69 (2000), 127–131. 3. P. Kronenberg, P.K. Rastogi, P. Giaccari, and H.G. Limberger, Relative humidity sensor with optical fiber Bragg gratings, Optics Lett 27 (2002), 1385–1387. 4. A.P. Russell and K.S. Fletcher, Optical sensor for the determination of moisture, Analytica Chimica Acta 170 (1985), 209 –216. 5. R. Hypszer and H.J. Wierzba, Fiber optic technique for relative humidity sensors, Proc SPIE 3054 (1997), 145–150. 6. D.C. Bownass, J.S. Barton, and J.D.C. Jones, Detection of high humidity by optical fibre sensing at telecommunications wavelengths, Optics Commun 146 (1998), 90 –94. 7. S. Otsuki, K. Adachi, and T. Taguchi, A novel fiber-optic gas-sensing configuration using extremely curved optical fibers and an attempt for optical humidity detection, Sensors and Actuators B 53 (1998), 91–96. 8. R. Jindal, S. Tao, J.P. Singh, and P.S. Gaikwad, High dynamic range fiber optic relative humidity sensor, Optical Eng 41 (2002), 1093– 1096. 9. S. Muto, O. Suzuki, T. Amano, and M. Morisawa, A plastic optical fibre sensor for real-time humidity monitoring, Measurement Sci and Technol 14 (2003), 746 –750. 10. P. Gaikwad, Chemically deposited optical fiber humidity sensor, M.Sc. thesis, Mississippi State University, 2003. © 2004 Wiley Periodicals, Inc. BEAM SWITCHING AND SCATTERING ANALYSIS FOR A MICROWAVE HOLOGRAM MODELED BY DISK LATTICES WITH TRANSVERSELY MODULATED SIZES Wenhao Zhu,1 Derek A. McNamara,1 and Jafar Shaker2 School of Information Technology and Engineering University of Ottawa 800 King Edward Street Ottawa, Ontario, K1N 6N5, Canada 2 Communication Research Center Canada 3701 Carling Ave. Ottawa, Ontario, Canada 1 Received 8 May 2004 ABSTRACT: The scattering from a microwave volume hologram made of layered 2D printed disk lattices is analyzed based on the dynamic interaction theory between small metallic disks and fields. The planar disk lattice has its disk size varying periodically in one direction (x) while remaining uniform in the other direction (y) to simulate a 1D holographic grating. The transmitted and reflected fields under TE planewave incidence are obtained in closed form. Beam switching can be achieved with a volume hologram thus implemented and the switching angle can be controlled by the slant angle of the grating with respect to the hologram’s normal. © 2004 Wiley Periodicals, Inc. Microwave Opt Technol Lett 43: 390 –394, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop.20480 Key words: thick-volume hologram; artificial dielectric; printed disk lattices; dynamic interaction fields 1. INTRODUCTION Microwave extension of the optical volume hologram concept is of practical interest, but has rarely been explored so far. One of the challenges in realization of microwave holography is how to form interference patterns in the refractive index of a dielectric used in the microwave-frequency band. Recently, artificial dielectric technology has been used to realize a thick-volume hologram operating in the microwave frequencies [1]. The procedure uses the low-cost, well-developed printed-circuit process to fabricate planar lattices of conducting “patches” on real dielectric layers, and then cascade these layers to form a 3D crystal structure. The effective-medium theory [2] can be used in the design of such artificial dielectrics or volume holograms to determine the patch dimensions and the lattice parameters in order to obtain the required effective permittivity. For grating and volume hologram applications, artificial dielectrics are often required to have periodic modulations in the effective permittivity along certain directions within the lattice plane. This can be achieved in a PCB implementation by varying the patch sizes periodically along the corresponding direction in the planar lattices. To predict the scattering characteristics of such volume holograms, one can adapt coupled-mode theories [3] to the effective continuous grating medium, or apply a periodic structure method of moments (MoM) formulation [4] to the original discrete patch-lattice structure. The former clearly relies on how good the equivalence between the arrayed planar patch lattices and the continuous grating medium will be, while the latter is a numerical procedure and is physically less insightful. This paper briefly describes an analytical method for analyzing the scattered fields from a finite array of 2D planar disk lattices. The procedure is based on the dynamic interaction theory between Contract grant sponsors: Ontario Graduate Scholarship and NSERC 390 MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 43, No. 5, December 5 2004