The humidity sensors constructed from polymer optical fiber Bragg gratings (POFBG) respond to the water content change in the fiber induced by varying environmental condition. The water content change is a diffusion process. Therefore the response time of the POFBG sensor strongly depends on the geometry and size of the fiber. In this work we investigate the use of laser micromachining of D-shaped and slotted structures to improve the response time of polymer fiber grating based humidity sensors. A significant improvement in the response time has been achieved in laser micromachined D-shaped POFBG humidity sensors. The slotted geometry allows water rapid access to the core region but this does not of itself improve response time due to the slow expansion of the bulk of the cladding. We show that by straining the slotted sensor, the expansion component can be removed resulting in the response time being determined only by the more rapid, water induced change in core refractive index. In this way the response time is reduced by a factor of 2.5.
© 2015 Optical Society of America
The physical and chemical properties of polymeric materials are rather different to silica, potentially making them attractive for researchers to exploit in device and sensing applications. In particular, polymer optical fiber (POF) sensors can offer large strain limits, high fracture toughness, extensive flexibility in bending, large and negative thermo-optic coefficients and for some materials an affinity for water. In addition, polymers have excellent compatibility with organic materials, giving them great potential for biomedical applications . Recently, fiber Bragg gratings (FBGs) have been inscribed into step index  and microstructured polymer fibers  made of poly(methyl methacrylate) (PMMA) materials, TOPAS  and polystyrene polymer fibers . PMMA based optical fiber Bragg gratings (POFBGs) were found to be sensitive to relative humidity, which is in contrast to FBGs in glass fiber which do not show any intrinsic humidity sensitivity . The affinity for water of PMMA leads to a swelling of the POF and an increase of refractive index, both of which contribute to an increase in the Bragg wavelength of a POFBG written in the PMMA fiber . This is a potentially very useful property, which has possible applications in chemical processing, agriculture, food storage, paper manufacturing, semiconductor and pharmaceutical industries, where humidity is monitored and controlled to ensure product quality. POFBGs have recently been applied to quantify the small amount of water dissolved in aviation fuel .
A typical POFBG humidity sensor is made of PMMA based optical fiber. A change of environmental conditions could lead to a change in the water content of the PMMA, which is a function of time . The process of water absorption or desorption in PMMA can be described by the diffusion theory of mass transfer. Although grating sensors made of different PMMA fibers may exhibit different response times and sensitivity, for the same type of PMMA fiber the humidity response time of POFBGs is determined by the fiber geometry . In our previous work , the response time of a POFBG was improved (lowered) by reducing the fiber diameter using chemical etching. Alternatively, a laterally accessible microstructured polymer fiber with a side-slotted cladding was suggested for further improvement in humidity time response [10, 11 ]. However, such slotted specialty fiber is not commercially available and also has the drawbacks of high loss and butt-coupling difficulty . Recently, advanced ultrafast lasers have opened up a new avenue for microfabrication due to the ultrashort pulse width and extremely high peak intensity, which enable spatially localized modification either on the surface or in the bulk of materials [12, 13 ]. In the last few years, different laser microfabrication methods have been explored, leading to microstructures with enhanced optical, chemical and biological properties [14, 15 ].
In this work we modify POFBGs by utilizing a laser micromachining technique. An excimer laser was employed to ablate the PMMA optical fiber and create microstructures (in the form of either a D-shape or a side-slot) in POFBGs. Both approaches are shown to improve the response time, but for the slotted device it is necessary to pre-strain the device to remove the contribution of the water induced expansion of the fiber cladding, which is relatively slow.
The Bragg wavelength of a PMMA-based POFBG depends on the core effective refractive index (RI) and the grating pitch, both of which are a function of temperature T and the water content w. For constant temperature, the Bragg wavelength change ΔλB of a POFBG against humidity can be expressed as :16]:Equation (2) can be used to define the water concentration in the polymer fiber for either water absorption (in-diffusion) or desorption (out-diffusion). In terms of normalized concentration, Eq. (2) can be rewritten for absorption:
From Eq. (1), one can see that there are two factors contributing to the wavelength change of a POFBG. One is the core effective RI change induced by the humidity change, i.e., the RI humidity dependence. The other is the volumetric change of the fiber induced by the moisture absorption. Since the fiber cladding, which is made of PMMA in this work, forms the largest portion of the fiber’s volume or weight, the volumetric change of fiber can be dominated and approximated by that of PMMA cladding. If a POFBG is fabricated in a fiber with smaller diameter it will result in faster response time, as indicated in Eq. (3). In our previous work different POFs were investigated by using chemical etching . Some POF (e.g., POF1 in ) shows a slow response time. Significant reduction of fiber diameter therefore is required to improve response time, which usually leads to unstable POFBG performance and a fragile structure. In this work we demonstrate improved POFBG performance by using laser micromachining.
3. POFBG fabrication and laser micromachining
Due to the high optical loss of POF in the 1550 nm wavelength region  a short length (7 cm) of POF is used to construct a POFBG sensor, which is connected to a standard single-mode silica fiber down-lead using UV curable glue. POF cannot be cleaved in the same fashion as silica fiber, neither can it be spliced using a fusion splicer. Hence, a UV gluing technique was developed  in which POF is cut by using hot blade to get a clean end face, then butt-coupled with an angled single-mode silica fiber pigtail; the butt-coupling joint is glued with UV curable glue (Norland 76). The PMMA step-index POF contained a 5 mm long FBG, fabricated by illuminating the fiber with 325 nm ultraviolet (UV) light from a HeCd laser through a phase mask placed on top of the POF. It should be pointed out that before grating fabrication the POF was annealed at 80°C over 7 hours. Therefore all the POFBGs were annealed. This is important because the mechanical property of polymer optical fiber is affected by the remaining stress in the fiber. This stress may vary with environmental conditions, generating variability in POFBG response. This effect exists in both step indexed and microstructured polymer fibers [17, 18 ].
Excimer lasers can produce light in the deep UV region of the spectrum and most polymer materials will absorb such radiation. Hence excimer laser micromachining has been used as an interesting approach to process polymeric materials. Laser micromachining occurs when the laser beam interacts with a material and its molecular bonds are broken by the absorption of the UV photons. The response of a material to the laser pulse will depend on the optical absorption and the photochemical, thermal and mechanical effects [12–14 ]. To produce a designed pattern and structure on a sample, an excimer laser incorporating the mask projection method has been used. Figure 1 schematically illustrates the excimer laser and mask scanning system.
The excimer laser system (S8000 Excimer Laser Micromachining Workstation, Exitech Ltd Oxford, UK) produces a 248 nm (KrF) laser beam and contains a high precision air-bearing work piece handling stage and XY mask stage. The laser beam is homogenized to produce a uniform intensity profile at the mask plane. The image of the mask (square aperture) is then projected through a high resolution demagnification lens (demagnification of × 10, NA of 0.3) onto the surface of a POFBG sample which is fixed on a micro-positioning workpiece stage (Aerotech Inc. USA) controlled by a computer to move in the XYZθ directions. The final feature to be produced on the fiber sample is therefore ten times less than the mask size. The laser micromachined microstructures, Figs. 1(b) and 1(c), can be achieved by moving the workpiece sample under the static aperture laser beam and triggering the laser output, firing with a defined incremental distance movement of workpiece stage along the fiber axis direction. The slotted structure was formed with the laser beam size (mask image) smaller than the fiber diameter while the D-shape was created with the beam size larger than fiber diameter. For example, using a mask of 400 µm × 400 µm in size, a laser beam of 40 μm × 40 μm can be projected onto the POFBG sample surface and hence a slot with width of 40 μm will be produced.
Multiple scan with laser radiation fluences of 1.37 J/cm2 and different masks have been used to generate the designed microstructures. Laser micromachined structures (Fig. 2 ) varying in width, ranging from a narrow slot, an open bay to a D-shape, can be achieved by the use of different masks with a size of 400 μm, 800 μm and 1.6 mm, respectively.
4. Experiments and results
Experiments were carried out to study the humidity response of the micromachined POFBGs. In our work, four POFBG samples (listed in Table 1 ) with cladding diameter of 130 µm, FBG length of 5 mm, and laser micromachined length of 5 mm (the same as FBG length) were used for humidity sensing experiments: one was normal POFBG at 1536 nm without any modification; one was slotted POFBG at 1563 nm with slotted-depth of 33.1 µm, and other two were D-shaped POFBGs at 1572 nm with depths of 44.5 µm (‘D-shaped 1’) and 50.5 µm (‘D-shaped 2’).
The POFBGs were placed inside an environmental chamber (Sanyo Gallenkamp) to allow operation in a controlled temperature and humidity environment. They were illuminated via a fiber circulator with light from a broadband source (Thorlabs ASE730) and the wavelengths of the POFBGs were monitored by using an IBSEN I-MON 400 wavelength interrogation system as shown in Fig. 3 . The environmental chamber was set at constant temperature (24°C) and programmed to change the relative humidity (RH) varying from 40% to 90% with an increment of 10% RH. The wavelength responses of the four POFBGs with humidity change were recorded and plotted in Fig. 4(a) where the wavelength traces for the four POFBGs were offset to give a better view. The step-changed RH in environmental chamber was measured by using a built-in RH sensor and is plotted in Fig. 4(a) as well. To fully observe the time response of POFBGs the dwelling time at each relative humidity level was set to 90 minutes. One can clearly see that it takes a very short time for the chamber to establish the relative humidity to the setting value; however, for a normal POFBG it takes a much longer time to reach a stable response. The micromachined POFBGs show different improvements in time responses.
To compare the performances of the different POFBG humidity sensors, the measured humidity responses when the relative humidity was step-changed from 60% RH to 70% RH are shown in Fig. 4(b). To simplify the comparison of the sensors which have different Bragg wavelengths, we chose to plot the time response of the relative wavelength change, Δλ/Δλmax, which is the ratio of the grating’s Bragg wavelength deviation from its original value to the maximum wavelength deviation induced by the humidity change . The response time then was estimated as the time of relative wavelength change being increased to 90% for the humidity step rise. For the normal POFBG sample (without any laser micromachining), the response time was estimated as 54 min, which was in agreement with 53 min reported in our previous work . The measured result for the slotted POFBG with a depth of 33.1 µm does not show much improvement on response time. Its response almost overlapped with that of the normal POFBG. We then tried to remove more cladding of the fiber. Two D-shaped POFBGs were machined, which showed a much improved response time, estimated as 30 min and 24 min for ‘D-shaped 1’ and ‘D-shaped 2’, respectively.
From the results it looks like the greater the volume of cladding removed, the faster the response of the POFBG. According to Eq. (1), the wavelength change of POFBG induced by environmental humidity consists of two parts: humidity dependent refractive index change and humidity induced volumetric change in the fiber core. Thus there are two ways to improve the response time of a POFBG humidity sensor. Firstly if the moisture can reach the fiber core in a short time it means a fast response time. Secondly, reducing fiber size would accelerate the moisture diffusion process, as suggested by Eq. (3) and Eq. (4). The slotted POFBG has a depth of 33.1 µm, which does not reach the fiber core but allows the moisture into the fiber core more quickly than a normal POFBG. Examining the wavelength response of the slotted POFBG in Fig. 4(b) reveals a quick change at the start, then it drops a little bit and eventually it rises and overlaps with that of the normal POFBG. This indicates that the moisture does reach the fiber core more quickly, producing a quick change of Bragg wavelength in the initial stage as the core index changes. However, the volumetric change in the fiber core not only depends on the moisture reaching the core but also on the volumetric change in the fiber cladding. Since the volume of fiber cladding is much larger than that of fiber core, the humidity induced volumetric change is dominated by the fiber cladding response time. From the slotted POFBG response one can see that there is a kink after the quick start. This kink also exists in all the machined POFBG responses to different extents. We believe it arises from the time lag between the humidity induced refractive index change in the fiber core and the volumetric change contribution from the bulk of the fiber cladding. This explains why the response of the slotted POFBG is almost overlapped with that of the normal POFBG after the kink.
In the case of the D-shaped POFBGs, the laser micromachining takes off a significant part of the POFBG cladding and so the volumetric change of the cladding takes less time to complete. Consequently, the time lag is smaller than that of the slotted POFBG response, and overall response time is much improved, showing a smaller kink. The response time of the deeply machined POFBG (‘D-shaped 2’) exhibits the fastest response time of 24 min, showing great improvement over the normal POFBG.
The above analysis can be verified by further experiment. As aforementioned, both humidity dependent refractive index change and humidity induced volumetric change in the fiber core contribute to the POFBG wavelength change. If the quick wavelength change in the initial stage is induced by humidity dependent refractive index and overall response time is dominated by the volumetric change of fiber cladding, removing the volumetric change will accelerate the time response of the POFBG. We proposed a technique in  to remove the humidity induced volumetric change in a POFBG by pre-straining the POFBG. The POFBG is strained by 2000 µɛ using a translation stage and then glued to an INVAR bar. The length of the PMMA optical fiber between the two anchoring points does not vary with PMMA swelling (given that the applied strain was larger than any humidity induced fiber swelling). The device length change due to changing temperature in this case is not determined by the POF thermal expansion but by the INVAR (the influence of the glue is negligible as the glued points are very small). Due to the low linear thermal expansion of INVAR (1.2 × 10−6 /°C) the pre-strain applied may vary with a rate of only 1.2 µɛ/°C. However, in any case during the experiment the thermal change is negligible as temperature is fixed in the environmental chamber. With this arrangement, the POFBG humidity responsivity only relies on the humidity induced refractive index change. The relative wavelength changes of the pre-strained slotted POFBG over 10%RH step change (from 60% RH to 70% RH) is plotted in Fig. 5 together with that when it is unstrained.
It can be seen that in the response of the pre-strained slotted-POFBG there is no kink which exists in the responses of both the slotted and D-shaped POFBGs, as shown in Fig. 4(b). Moreover, the response time of the pre-strained slotted-POFBG is around 20 min, which is reduced by a factor of 2.5 than that of the slotted POFBG with no strain applied. These two features clearly indicate that pre-straining can remove humidity induced volumetric change in the POFBG and the response time of a POFBG humidity sensor can be accelerated by this process. However, this acceleration is achieved at the cost of reduced POFBG humidity sensitivity as the contribution from humidity induced volumetric change is removed . It can be seen from Fig. 5 that 0.2 nm and 0.6 nm of wavelength change in pre-strained and unstrained slotted POFBGs are induced by 10% humidity step change, respectively. The humidity sensitivity of the slotted POFBG was 60 pm/%RH with no strain applied, and reduced to 20 pm/%RH when a pre-strain of 2000 μɛ applied. The reduction in sensitivity is not really problematic as current interrogation systems typically provide resolutions around 1 pm, readily enabling RH monitoring to better than 1%.
The response time of the pre-strained slotted-POFBG is even shorter than that of D-shaped POFBGs. Therefore we expect the response time of pre-strained D-shaped POFBGs can be further reduced. However, the optical performance of a pre-strained D-shaped POFBG is not as stable as that of the pre-strained slotted-POFBG. When we applied strain to the D-shaped POFBG, the reflectivity dropped rapidly and then vanished, even though the fiber was not broken. Removing the strain did not recover the grating response. The response times of all the POFBGs used in this work are summarized in Table 1.
The response time of a POFBG to humidity changes is mainly influenced by the geometry and size of the fiber. Laser micromachining can provide a flexible method to modify POFBG sensor performance. In this work we have fabricated the slotted and D-shaped structures over POFBGs. The D-shaped structure provides an improved response time, similar to the existing technique of fiber etching. Unfortunately, fibers heavily machined in this way become difficult to handle and lack robustness, similar to very small diameter etched fibers. The slotted fiber provides a way of allowing water in the environment rapid access to the fiber core, whilst preserving the strength of the device. We have discovered that this geometry does not of itself significantly improve response time due to the dominant contribution of the swelling of the fiber cladding material. We have shown that by pre-straining the sensor, this contribution can be removed allowing a much faster response due to the core index variation (a factor of 2.5 in our experiment). The penalty for this approach is a reduction in sensitivity, but the magnitude of the wavelength shift observed is still large enough to allow sub-1% relative humidity resolution with typical interrogation systems.
The authors would like to acknowledge the support from the European Union Seventh Framework Programme under grant agreement No. 314032 and No. PIRSES-GA-2013-612267, and the project of Sêr Cymru NRN097.
References and links
1. D. J. Webb and K. Kalli, “Polymer Fiber Bragg Gratings,” in Fiber Bragg Grating Sensors: Thirty Years From Research to Market, A. Cusano, A. Cutolo, and J. Albert ed. (Bentham Science, 2010).
2. Z. Xiong, G.-D. Peng, B. Wu, and P. L. Chu, “Highly tunable Bragg gratings in single-mode polymer optical fibers,” IEEE Photonics Technol. Lett. 11(3), 352–354 (1999). [CrossRef]
3. H. Dobb, D. J. Webb, K. Kalli, A. Argyros, M. C. J. Large, and M. A. van Eijkelenborg, “Continuous wave ultraviolet light-induced fiber Bragg gratings in few- and single-mode microstructured polymer optical fibers,” Opt. Lett. 30(24), 3296–3298 (2005). [CrossRef] [PubMed]
4. W. Yuan, L. Khan, D. J. Webb, K. Kalli, H. K. Rasmussen, A. Stefani, and O. Bang, “Humidity insensitive TOPAS polymer fiber Bragg grating sensor,” Opt. Express 19(20), 19731–19739 (2011). [CrossRef] [PubMed]
5. N. G. Harbach, “Fiber Bragg gratings in polymer optical fibers,” PhD Thesis, Lausanne, EPFL (2008).
6. C. Zhang, W. Zhang, D. J. Webb, and G.-D. Peng, “Optical fibre temperature and humidity sensor,” Electron. Lett. 46(9), 643–644 (2010). [CrossRef]
7. C. Zhang, X. Chen, D. J. Webb, and G.-D. Peng, “Water detection in jet fuel using a polymer optical fibre Bragg grating,” Proc. SPIE 7503, 750380 (2009). [CrossRef]
9. W. Zhang, D. J. Webb, and G.-D. Peng, “Investigation into time response of polymer fiber Bragg grating based humidity sensors,” J. Lightwave Technol. 30(8), 1090–1096 (2012). [CrossRef]
12. P. E. Dyer, “Excimer laser polymer ablation: twenty years on,” Appl. Phys., A Mater. Sci. Process. 77, 167–173 (2003).
13. A. S. Holmes, “Excimer laser micromachining with half-tone masks for the fabrication of 3-D microstructures,” Science, Measurement and Technology, IEE Proc. 151, IET, (2004). [CrossRef]
14. J. Burt, A. Goater, A. Menachery, R. Pethig, and N. Rizvi, “Development of microtitre plates for electrokinetic assays,” J. Micromech. Microeng. 17(2), 250–257 (2007). [CrossRef]
15. G. N. Smith, T. Allsop, K. Kalli, C. Koutsides, R. Neal, K. Sugden, P. Culverhouse, and I. Bennion, “Characterisation and performance of a Terfenol-D coated femtosecond laser inscribed optical fibre Bragg sensor with a laser ablated microslot for the detection of static magnetic fields,” Opt. Express 19(1), 363–370 (2011). [CrossRef] [PubMed]
16. J. Crank, The Mathematics of Diffusion (Oxford University, 1975).
17. K. E. Carroll, C. Zhang, D. J. Webb, K. Kalli, A. Argyros, and M. C. J. Large, “Thermal response of Bragg gratings in PMMA microstructured optical fibers,” Opt. Express 15(14), 8844–8850 (2007). [CrossRef] [PubMed]
18. W. Yuan, A. Stefani, M. Bache, T. Jacobsen, B. Rose, N. Herholdt-Rasmussen, F. K. Nielsen, S. Andresen, O. B. Sørensen, K. S. Hansen, and O. Bang, “Improved thermal and strain performance of annealed polymer optical fiber Bragg gratings,” Opt. Commun. 284(1), 176–182 (2011). [CrossRef]