We report the effect of coating thickness on the sensitivity of a relative humidity (RH) sensor based on an Agarose coated photonic crystal fiber interferometer for the first time. An experimental method is demonstrated to select an optimum coating thickness to achieve the highest sensitivity for a given RH sensing range. It is shown that the Refractive Index (RI) of the coating experienced by the mode interacting with the coating depends on the thickness of the coating. It is observed that the spectral shift of the interferometer depends on both the bulk RI change and the thickness change of the Agarose coating with respect to an RH change. The RH sensitivity of the sensor has a significant dependence on the thickness of the coating and the sensor with highest sensitivity shows a linear response for RH change in the range of 40-90% RH with a humidity resolution of 0.07%RH and a fast response time of 75 ms for an RH change from 50% to 90%.
© 2013 OSA
Humidity sensors are necessary in many areas, such as meteorological services, the chemical and food processing industry, civil engineering, air-conditioning, horticulture, electronic processing, human breath rate monitoring, paper and textile production and many other fields. Optical fiber humidity sensors, by comparison to their electronic counterparts, offer specific advantages such as small size and low weight, immunity to electromagnetic interference, corrosion resistance and the potential for remote operation. Most fiber optic humidity sensors work on the basis of a hygroscopic polymer material coated over the optical fiber to modulate the light propagating through the fiber [1–4].
Recently photonic crystal fiber (PCF) interferometers based on microhole collapse have taken on increased importance in sensing applications [5–13] due to their fabrication simplicity which involves only cleaving and splicing operations. The different configurations reported so far are the PCF with two collapsed regions separated by a few centimeters , a stub of PCF with the cleaved end fusion spliced to the distal end of a standard single mode fiber [6–9] and a short section of PCF longitudinally sandwiched between single mode fibers [10–13]. The advantage of the last two configurations is that the modal properties of the PCF are exploited but conventional optical fibers are used to connect to the interrogation system, thus leading to more cost-effective interferometers. The photonic crystal fiber interferometer (PCFI) configuration in which the two ends of a PCF are fusion spliced to lead-in and lead-out single mode fibers has already been demonstrated for strain [10,11] and refractive index (RI) [12,13] sensing. We have recently demonstrated for the first time  a humidity sensor based on this type of PCFI configuration using an Agarose coating. The RH sensitivity of the sensor was 8 pm/%RH in the range 30-80%RH. The aim of this work is to study the effect of the coating thickness of Agarose so as to improve the sensitivity of this type of RH sensor. To achieve this several PCFI devices with the same length but different coating thicknesses are fabricated and their individual humidity response is studied. We find that the spectral shift of a PCFI device depends on both the bulk RI change and the thickness change of the Agarose coating with respect to RH change. Also we find that the RH sensitivity of the sensor is significantly influenced by the thickness of the coating. The sensor with the highest sensitivity shows a humidity sensitivity of 137 pm/ %RH in the range 40-90%RH. The work carried out in this paper also can be applied to sensors based on optical fibers coated with other materials for the selection of a suitable operating point in terms of sensitivity and range of operation.
2. Experimental investigation and discussion
The transmission type PCFI in our experiment is fabricated by fusion splicing a length of commercial PCF (LMA-10, NKT Photonics) between standard optical fibers with a conventional splicing machine. The splicing was carried out in such a way that the voids of the PCF collapsed completely over a short region, typically 200 μm long, close to both the splice points. The collapsed region itself enables the excitation of two modes in the PCF. The spectrum of the interferometer depends on the splicing conditions which in turn can be controlled by the arc time and duration . To interrogate the interferometer light from a broadband source, SLED (COVEGA Corporation, SLD1005) is launched into the device, and the transmitted light is fed to an optical spectrum analyzer (OSA), see Fig. 1(a) . Compared to the reflection type interferometer we previously reported for humidity sensing [7–9], in the present approach a significant advantage is that there is no possibility of any contaminants such as dust entering the voids of the PCF after fabrication. In addition, since the device is a transmission type sensor, it does not require the use of a fiber optic circulator. To understand how this interferometer works it is useful to analyze the guided beam when it travels from the SMF to the PCF (Fig. 1(b)). The fundamental SMF mode begins to diffract when it enters the collapsed PCF region, which is solid glass. Because of diffraction the mode broadens allowing the excitation of core and cladding modes in the PCF section [6, 12, 13]. Since the modes propagate with different phase velocities, then over a length of PCF the modes accumulate a differential phase shift. At the end of the PCF, the modes reach the second solid glass section, i.e., the other collapsed end of the PCF. They will thus further diffract and will be recombined through the spatial filtering of the output SMF. Because of interference, the optical power transmitted by the device will be maximum at certain wavelengths and a minimum at others since the phase difference accumulated over the PCF section is wavelength dependent.
In our experiment in order to study the effect of coating thickness on the sensitivity of an RH sensor based on an Agarose coated PCFI we fabricated four identical PCFI devices with a PCF length of 40 ± 0.015 mm and fringe spacing of 14.8 ± 0.22 nm. The transmission spectra of these PCFIs are shown in Fig. 2(a) . The similar spectra obtained for different devices show the very good fabrication repeatability for the interferometer. The response of these PCFIs to external RI variations is studied by immersing the devices in different calibrated RI solutions, over an RI range from 1.33 to 1.45. It is observed that for the same change in RI, the shift in the individual peaks of the interference pattern varies slightly. To deal with these slight variations in determining sensitivity, average peak shift is calculated at each RI value of the sensor response, over the wavelength range observed. Figure 2(b) shows the average spectral peak shift of the device with respect to external RI. As expected all the devices showed very similar RI responses. It is also observed that as expected the RI sensitivity of the interferometer is higher in the region close to the RI of the PCF material - silica (1.44).
Coating with Agarose is carried out by drawing the interferometer through a hot (65 °C) Agarose solution . A schematic diagram of the experimental setup for Agarose coating is shown in Fig. 1(a). The solution is prepared by dissolving 1 wt% Agarose in distilled water. In order to undertake the coating process the fiber is fixed straight and horizontally above a translation stage. Below the fiber a heater is fixed on a translation stage. A small container placed at the top of the heater is filled to the rim of the container with a hot Agarose solution. Because of surface tension the surface of the solution forms a dome-like shape which projects slightly above the rim of the container. The position of this container can be adjusted to allow the fiber to pass through this dome of Agarose solution. The temperature of the heater is set at 65 °C. The fiber is drawn through the hot Agarose solution using a translation stage which is software controlled using a computer. This arrangement allows for good repeatability of the coating parameters.
A desired coating thickness can be achieved using our setup by varying the drawing speed of the fiber through the solution or by passing the fiber multiple times through the solution. Practically the latter technique is found to be best in order to achieve a repeatable thickness. In our experiment the fiber is drawn through the solution multiple times but with a constant speed of 5 mm/sec. Accurate measurement of the film thickness during the coating process is difficult due to the fiber geometry and thus in our experiment the device’s spectrum is monitored each time the fiber passes through the solution. The observed shift in the interference spectrum confirms the formation of a coating on the PCFI as it is in agreement with the expected red shift of the PCFI response with an increase in ambient RI for the interferometer (Fig. 2(b)). In our experiment after passing the PCFI through the solution for the first time an average spectral red shift in the range of 1.2-2.5 nm is observed in the transmission spectrum of the device compared with its initial spectrum. On passing the device through the solution for the second time this shift increases to a range of 3.0-7.5 nm and so on.
The four PCFI devices are coated with different thickness of Agarose film by passing the fiber through the solution multiple times and simultaneously comparing the spectral shift of the coated device with its RI response, thus ensuring that each PCFI is coated with an Agarose film with a different RI, produced by different film thicknesses. In the case of device A the coating process is stopped when the average peak shift reaches a value of 2.3 nm and for devices B, C and D the coating process is stopped when the shift is 3.55 nm, 14.5 nm and 26.6 nm respectively. A schematic diagram of the Agarose coated PCFI (AC-PCFI) and a drawing of the cross section of the PCF employed are shown in Fig. 1(b).
The parameters of the AC-PCFI devices are listed in Table 1 . The peak shift is calculated as the average difference in the peak positions for all the individual peaks in the transmission spectrum of the device before and after coating it with Agarose. The effective refractive index of the coating is calculated by comparing the peak shift obtained for a particular coating with the peak shift data obtained for the RI response of that device. The thickness of the coated device is estimated using an optical microscope at a room RH of 60 ± 2%. From the data shown in the Table 1 it is obvious that when the coating thickness increases, the effective RI of the coating experienced by the cladding mode of the PCFI increases. Therefore control of the RI of the coating of the PCF interferometer can be realized by selecting a suitable thickness for the coating.
The RH responses of the AC-PCFI devices are studied by placing them inside a controlled environmental chamber (Electro Tech Systems inc., Model 5503-00 with Package F). The ambient temperature during the study was set at 24 ± 1 °C at normal atmospheric pressure. In our devices the interaction of the coating is solely with the cladding modes, since the core mode is isolated from the external environment. The interaction of the cladding modes with the Agarose coating changes the effective index of the cladding mode and consequently the phase difference between the cladding mode and the core mode. As a result the interference pattern shifts, that is the position of the interference peaks and valleys change. An increase in the effective RI of the cladding mode causes a spectral red shift and a decrease in the effective RI causes a blue shift.
The previous studies on humidity sensing where an Agarose coated optical device is used as the sensor head considered only the bulk RI change of the coating induced by changes in RH. However there exists some conflicting evidence regarding the refractive index change of an Agarose coating with respect to RH [3, 4, 9, 14–18]. Our present study in this paper gives further insight into the behavior of an Agarose coating when it’s ambient RH changes.
We have studied the thickness change of the Agarose coating for a change of RH. Figure 3 shows the thickness of the Agarose coating for different PCFI devices at room RH of 60 ± 2% and at a higher RH of >90% estimated using an optical microscope. It is observed that the thickness of the coating increases significantly (>60%) when RH increases from ~60% to >90% RH. So we conclude that apart from the RI change the thickness of the Agarose coating also increases when RH increases. The swelling characteristic of Agarose is also reported in . When the RH level increases, more water molecules are diffused into the Agarose coating, resulting in the inflation of the Agarose and increase in the thickness of the coating. Similar to any other swelling polymer an increase in water content will decrease the bulk refractive index of the Agarose coating according to the Lorenz- Lorentz relation . Also it has been shown in [13,20] that for fiber devices with coatings of sub-micron thickness as the thickness of the coating increases the effective RI experienced by the cladding mode increases even though the bulk RI of the coating is a constant. This effect is also confirmed by Table 1 where a spectral red shift observed when the thickness of the coating increased is tabulated, even though the local RH is constant. Thus when RH changes two factors alter the effective RI of the cladding mode; the bulk RI change of the coating and the thickness change of the coating. When the RH increases the decrease in the bulk RI of the coating causes a blue shift to the interference spectrum while the increase in the thickness of the coating causes a red shift to the interference spectrum.
We find that for coatings of thickness less than the reach of the evanescent wave part of the cladding mode, which is normally significant up to one third the wavelength of the propagating light [21,22]; the increase in the effective RI (as a result of the increase in the thickness of the Agarose) dominates and is more significant than the effect of the bulk RI change, resulting in a net red shift of the interference spectrum for all RH values and this is the case with the devices A and B in our experiment. Figure 4(a) shows the average peak shift of the interference pattern of the AC-PCFI A and B with respect to RH. The peak wavelength shift shown is normalized with respect to the peak value at the lowest measured RH (25% RH). The AC-PCFI A shows a spectral red shift with an increase in RH and the observed shift is linear in the range from 25 to 90% RH with a slope of 5 pm/%RH. Above 90% RH the sensitivity of the device is much higher most likely due to the water vapor condensation on the coating at these higher RH values. With an increase in humidity the AC-PCFI B shows a red shift with a slope of 9 pm/%RH in the range 25 to 90% RH. It can be concluded that when the thickness of the coating increases the humidity sensitivity of the AC-PCFI also increases. One possible reason for this behavior is that when the thickness of the coating increases, the effective RI of the coating then lies in the high RI sensitivity region of the PCFI (the higher slope region in Fig. 2(b)) and hence the change in the effective RI of the coating with respect to an RH change results in an increased spectral shift.
However when the coating thickness is greater than the reach of the evanescent wave part of the cladding mode, then for an increase in the RH the influence of an RI change resulting from a thickness change has less effect compared to the bulk RI change of the coating. In such cases the effect of bulk RI dominates the effective RI experienced by the cladding mode. In our experiment it is observed that when the coating thickness is greater than ~800 nm, which is the case with AC-PCFI devices C and D, the observed shift is blue for the interference spectrum of the AC-PCFI device when the RH increases. For verification we have calculated the evanescent wave penetration depth of device C at 60% RH using Eq. (1) of reference  by setting the value of the RI of the fiber material silica, n1 = 1.44, the RI of the Agarose coating, n2 = 1.408 (Table 1) and assuming the angle of incidence at the fiber–coating interface θ = 90°. Using these values the calculated penetration depth is 817 nm, verifying the observation that the wavelength shift changes from a red to a blue shift when the coating thickness is in the region of 800 nm.
The humidity responses of the AC-PCFI devices C and D are shown in Fig. 4(b). The AC-PCFI C shows a spectral red shift when humidity increases from 25% RH to 60% RH and then it shows a blue shift on a further increase of RH from 60% to 98%. This is because below 60% RH the thickness change factor dominates by comparison to the bulk RI change of the coating and above 60% RH the coating thickness is greater than the penetration depth of the evanescent wave portion of the cladding mode interacting with the coating so that the effective RI of the cladding mode is mainly determined by the bulk RI of the coating. For AC-PCFI D the peak wavelength phase change point (RH at which the red shift changes to blue shift) of the RH response curve shifts to lower RH of 40% which is expected because here the coating is thicker than for AC-PCFI C. For AC-PCFI D since the RI of the coating is in a more RI sensitive region of the PCFI and the RH sensitivity observed is also higher compared to other devices with a smaller thickness of Agarose coating. The AC-PCFI D shows a sensitivity of 64 pm/%RH in the region 25-40% RH and 137 pm/%RH in the range 40-90%RH. Assuming the OSA used for measurement has a wavelength resolution of 0.01 nm, the AC-PCFI D has a humidity resolution of 0.07% RH in the RH range from 40% to 90% RH.
The calculated RH sensitivities of the AC-PCFI devices in different linear RH regions are listed in Table 2 , where the positive sensitivity values represent spectral red shift and negative values represent a blue shift. Above 90% RH the response is nonlinear so the Table 2 shows the average sensitivity.
A change of humidity inside the environment chamber from an initial value to a final value and back requires several minutes. The response time of Agarose coated PCFI sensor itself is observed to be much faster. Because of this changing the RH using the chamber is not a suitable means for studying the response time of the sensor. Instead we applied a step change in humidity to the sensor by directing a human breath exhale to the sensor. The resultant time dependant response is found to be faster than time response of the OSA. Therefore we used an optical power meter (PX Instrument Technology, PX2000-306) and a suitable wavelength output of a tunable laser source (Anritsu, Tunics plus CL/WB) to measure the change in the transmission of the device in response to a sudden RH change. As an example, the time dependant response of AC-PCFI device D is shown in Fig. 5 . The ambient humidity during the study was ~50%RH and the temperature ~23 °C. The estimated response time (10% base line to 90% signal maximum) of the sensor is about 75 ms, when RH jumps from 50 to >90%. The recovery time of a humidity sensor depends on how fast the water vapor is removed from the sensor which is proportional to the air flow surrounding the sensor. The estimated recovery time (90% signal maximum to 10% baseline) of the sensor is 110 ms, which decreases if a flow of dry air surrounds the sensor.
As the RH value is extracted from wavelength measurements, the sensor is free from any errors due to power fluctuations in the optical source. We have studied the RH response of the AC-PCFI devices four weeks after the initial study and it is observed that the devices shows a good repeatability for its RH response confirming the potential for long term stability for the sensor. The previous studies on humidity sensing where an Agarose coated optical fiber device is used as the sensor head do not give a clear picture of the refractive index change and the thickness change of Agarose coating with respect to RH. Those works considered the RI change of the coating only but the results presented in this paper reveal that there is a significant and complex dependence of the sensor response on the thickness of the coating. This result can potentially be applied to different types of coatings on a PCFI device that may result in improved sensor performance and opens an area for further research for this type of sensor head.
In conclusion we have coated a PCFI with Agarose layers of different thickness and demonstrated that the RI of the coating experienced by the mode interacting with the coating depends on the thickness of the coating. We also demonstrated that both bulk RI change and the thickness change of the Agarose coating with respect to RH affect the spectrum of the PCFI. We studied the effect of coating thickness on the RH response of the AC-PCFI devices and showed that the RH sensitivity of an AC-PCFI depends strongly on the thickness of the coating. The sensor with the highest sensitivity shows a linear response for RH change in the range of 40-90%RH with a humidity resolution of 0.07%RH. The response time of the sensor is 75 ms for an RH change from 50% to 90%. This work also provides the basis for the selection of an optimal operating point in terms of sensitivity and range of operation in the case of a PCFI coated with other materials for different sensing applications.
References and links
1. T. L. Yeo, T. Sun, and K. T. V. Grattan, “Fibre-optic sensor technologies for humidity and moisture measurement,” Sens. Actuators A Phys. 144(2), 280–295 (2008). [CrossRef]
2. J. Mathew, Y. Semenova, G. Rajan, P. Wang, and G. Farrell, “Improving the sensitivity of a humidity sensor based on fiber bend coated with a hygroscopic coating,” Opt. Laser Technol. 43(7), 1301–1305 (2011). [CrossRef]
3. J. Mathew, Y. Semenova, and G. Farrell, “A fiber bend based humidity sensor with a wide linear range and fast measurement speed,” Sens. Actuators A Phys. 174, 47–51 (2012). [CrossRef]
4. J. Mathew, K. J. Thomas, V. P. N. Nampoori, and P. Radhakrishnan, “A comparative study of fiber optic humidity sensors based on chitosan and agarose,” Sens. Transducers J. 84, 1633–1640 (2007).
6. R. Jha, J. Villatoro, and G. Badenes, “Ultrastable in reflection photonic crystal fiber modal interferometer for accurate refractive index sensing,” Appl. Phys. Lett. 93(19), 191106 (2008). [CrossRef]
7. J. Mathew, Y. Semenova, G. Rajan, and G. Farrell, “Humidity sensor based on photonic crystal fibre interferometer,” Electron. Lett. 46(19), 1341–1343 (2010). [CrossRef]
8. J. Mathew, Y. Semenova, and G. Farrell, “Photonic crystal fiber interferometer for dew detection,” J. Lightwave Technol. 30(8), 1150–1155 (2012). [CrossRef]
9. J. Mathew, Y. Semenova, and G. Farrell, “Relative humidity sensor based on an agarose infiltrated photonic crystal fiber interferometer,” IEEE J. Sel. Top. Quantum Electron. 18(5), 1553–1559 (2012). [CrossRef]
10. J. Villatoro, V. Finazzi, V. P. Minkovich, V. Pruneri, and G. Badenes, “Temperature-insensitive photonic crystal fiber interferometer for absolute strain sensing,” Appl. Phys. Lett. 91(9), 091109 (2007). [CrossRef]
11. D. Barrera, J. Villatoro, V. P. Finazzi, G. A. Cardenas-Sevilla, V. P. Minkovich, S. Sales, and V. Pruneri, “Low-loss photonic crystal fiber interferometers for sensor networks,” J. Lightwave Technol. 28(24), 3542–3547 (2010). [CrossRef]
13. M. Smietana, D. Brabant, W. J. Bock, P. Mikulic, and T. Eftimov, “Refractive-index sensing with inline core-cladding intermodal interferometer based on silicon nitride nano-coated photonic crystal fiber,” J. Lightwave Technol. 30(8), 1185–1189 (2012). [CrossRef]
14. J. Mathew, Y. Semenova, and G. Farrell, “Polymer coated photonics crystal fiber interferometer for relative humidity sensing,” A. Ghosh, and D. Choudhury (Eds.): Proc. IConTOP–II, pp. 73–78, (2011).
15. F. J. Arregui, Z. Ciaurriz, M. Oneca, and I. R. Matías, “An experimental study about hydrogels for the fabrication of optical fiber humidity sensors,” Sens. Actuators B Chem. 96(1-2), 165–172 (2003). [CrossRef]
16. K. J. Lee, D. Wawro, P. S. Priambodo, and R. Magnusson, “Agarose-gel based guided-mode resonance humidity sensor,” IEEE Sens. J. 7(3), 409–414 (2007). [CrossRef]
17. 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,” Sens. Actuators B Chem. 69(1-2), 127–131 (2000). [CrossRef]
18. M. Hernaez, C. R. Zamarreño, C. Fernandez-Valdivielso, I. del Villar, F. J. Arregui, and I. R. Matias, “Agarose optical fibre humidity sensor based on electromagnetic resonance in the infra-red region,” Phys. Status Solidi C 7(11-12), 2767–2769 (2010). [CrossRef]
19. J. J. Roberts, A. Earnshaw, V. L. Ferguson, and S. J. Bryant, “Comparative study of the viscoelastic mechanical behavior of agarose and poly(ethylene glycol) hydrogels,” J. Biomed. Mater. Res. B Appl. Biomater. 99B(1), 158–169 (2011). [CrossRef]
20. I. Del Villar, I. R. Matías, F. J. Arregui, and P. Lalanne, “Optimization of sensitivity in long period fiber gratings with overlay deposition,” Opt. Express 13(1), 56–69 (2005). [CrossRef]
21. J. Lu, Z. Chen, F. Pang, and T. Wang, “Theoretical analysis of fiber-optic evanescent wave sensors,” in Proceedings of IEEE Microwave Conference, (China-Japan Joint, 2008), pp.583–587, doi: 10.1109/CJMW.2008.4772500. [CrossRef]
22. S. T. Lee, P. S. Kumar, K. P. Unnikrishnan, V. P. N. Nampoori, C. P. G. Vallabhan, S. Sugunan, and P. Radhakrishnan, “Evanescent wave fibre optic sensors for trace analysis of Fe3+ in water,” Meas. Sci. Technol. 14(6), 858–861 (2003). [CrossRef]