A new kind of surface plasmon resonance (SPR) sensor based on silver-coated hollow fiber (HF) structure for the detection of liquids with high refractive index (RI) is presented. Liquid sensed medium with high RI is filled in the hollow core of the HF and its RI can be detected by measuring the transmission spectra of the HF SPR sensor. The designed sensors with different silver thicknesses are fabricated and the transmission spectra for filled liquids with different RI are measured to investigate the performances of the sensors. Theoretical analysis is also carried out to evaluate the performance. The simulation results agree well with the experimental results. Factors that might affect sensitivity and detection accuracy of the sensor are discussed. The highest sensitivity achieved is 6607nm/RIU, which is comparable to the sensitivities of the other reported fiber SPR sensors.
© 2013 Optical Society of America
Surface plasmon resonance (SPR) is a surface-sensitive analytical method for chemical and biochemical sensing [1, 2]. Since Kretschmann et al. released SPR phenomenon in the 1960s , sensors based on SPR have attracted considerable attention and been extensively studied. Most of these sensors adopt a prism based configuration that consists of a prism, a metal layer and the sensing layer [4, 5]. However, the prism based SPR sensors are usually bulky, difficult for remote sensing, and require expensive optical equipment. Optical fibers are perceived as being safer for in vivo use because of the optical signal and there is no electromagnetic interference. Thus many kinds of fiber SPR sensors were studied in recent years. Conventional solid core fiber SPR sensors [6–9] and fiber SPR sensors using fiber gratings, such as fiber Bragg grating (FBG), tilted FBG (TFBG), and long-period fiber grating (LPG) [10–12], are sensitive to the refractive index (RI) changes of surrounding sensed medium. Theoretical models of SPR sensors based on the structure of photonics crystal fiber (PCF) have also been released, and the sensed medium is outside the PCF as well [13–15]. Theoretically analysis are also made for fiber SPR sensors that can hold liquid sensed medium inside the central holes of the PCF [16–18] or micro-structured optical fiber (MOF) [14, 19], or the hollow core of the grating assisted fiber [20, 21]. However, because the bore size of the hollow cores in these sensors is as small as several microns, the deposition of the thin metal film on the inner wall of the hollow core is difficult. The gratings and complicated PCF or MOF structures make the fabrication even more difficult. We have not seen any report on the experimental investigation on the fiber SPR sensors holding the sensed medium in the hollow core of the fibers.
Owing to its simple structure and low loss properties in the visible and infrared regions, hollow fiber (HF) has been widely studied and is becoming one of the most commonly used optical fibers in many applications [22–25]. The structure of a silver coated hollow fiber usually consists of a supporting tube, a silver layer coated on the inner surface of the tube, and a hollow air core which is about several hundred microns in diameter. In most studies, HF is used in energy delivery for infrared or visible laser light. Therefore, the silver layer coated in HF should be as thick as several hundred nanometers to ensure a high reflation rate. Studies of utilizing such kind of energy transmitting HF in SPR sensing for the detection of liquids with high RI have not been reported.
In this paper we present a new kind of SPR sensor based on silver coated HF structure. Liquid sensed medium with high RI is filled in the hollow core of the HF and its RI can be detected by measuring the transmission spectra of the HF SPR sensor. The designed sensors with different silver layer thicknesses are fabricated and the transmission spectra for filled liquids with different RI are measured. Theoretical analysis based on the ray transmission model is also carried out to evaluate the performance of the sensor. Factors that affect sensitivity and detection accuracy of the sensor are investigated theoretically and experimentally. The presented sensor has potential applications in the oil industry or chemical industry where high refractive liquids such as benzene are often employed. For example, it can monitor the RI variation of liquid medium on-line by flowing the liquid through the sensor via a by-path.
2. Sensor configuration and theory
Figure 1 shows the structure of the silver coated HF. When the thickness of the silver layer is less than about 100 nm, it becomes a SPR sensor. Surface plasmon waves would be excited on the interface between the metal layer and the supporting tube when appropriate light transmits in the HF with high RI liquid medium in the hollow core. The SPR sensor can detect the RI of the filled sensed medium. Different from most fiber SPR sensors reported previously, the HF SPR sensor holds the liquid sensed medium inside the hollow core and the detection light is transmitted in the sensed medium. To satisfy the condition of total reflection, the liquid sensed medium should have a higher RI than the cladding of the fiber which is fused silica in our experiment. Therefore, the designed HF SPR sensor has an inherent advantage in the detection of liquids with high RI.
A ray transmission model for theoretically analyzing the performance of the HF SPR sensor is also illustrated in Fig. 1 [26, 27]. Input light is launched into the HF from a multi-mode fiber (MMF) and total reflects on the inner surface of the HF while passing through the HF SPR sensor. Here we only consider the meridional rays. While the surface plasmon waves are excited, part of the energy of the incident light is transferred to the evanescent surface waves, leading to an apparent decrease of reflectivity. The corresponding resonance condition for surface plasmons  is written as26]. The distribution of the input light power is close to Gaussian with profile expressed approximately as Fig. 1 follows Snell's Law .
The power of the output light of the HF is given as28]. The number of reflections K, in the fiber sensor area is a function of θ, core diameter D, and the length L, which is
Therefore, the generalized expression for the normalized transmittance of the HF SPR sensor can be expressed as
With Eq. (5), the performance of the HF SPR sensor could be numerically evaluated. Here we only consider the p-polarized light, however in the experiment the input light is the mixed light of both p-polarized and s-polarized lights. SPR effect demonstrates itself as a sharp minimum in the transmission spectrum. The wavelength where the resonance dip located is called resonance wavelength (RW). The RW will shift when the RI of the sensed medium changes. Therefore, by measuring RW from the transmission spectra of the HF SPR sensor, we can get the RI of the liquid sensed medium filled in the sensor.
HFs with different silver layer thicknesses are fabricated. A thin silver layer is coated on the inner surface of the flexible fused silica capillary with an inner diameter (ID) of 700μm by an improved liquid phase deposition method . The thickness of the silver layer should be less than 100 nm. Silver-coating on the inner wall with such a thin silver layer is difficult for the chemical liquid phase deposition method. The deposition time, flow rate of solutions and temperature must be carefully controlled to obtain a uniform and smooth silver layer which is also thin enough for SPR sensing. Silver nitrate and glucose solutions are used as the plating and reducing solutions respectively. The solutions are mixed and forced to flow through the glass capillary. Reduced silver particles adhere on the inner wall and gradually form a silver layer. The flow rate needs to be high to obtain good mixture of the solutions. Otherwise, the roughness of the silver layer increases. High temperature increases the deposition speed, which causes difficulties in controlling the silver layer thickness. However, low temperature leads bad silver layer quality. So the deposition temperature is 15°C in our experiments. Efforts have been made to smooth the silver layer surface. The pre-treatment to the glass surface by using SnCl2 solution gives a much smoother surface. The schematic diagram of the deposition method is shown in Fig. 2. A piece of fiber with a short length of about 5 cm is cut off from the original one meter long HF to be used as the sensor. The ID and length (D and L shown in Fig. 1) are 700 μm and 5 cm.
The measurements were performed on the experimental set-up illustrated in Fig. 3. The liquid sensed medium was injected into the sensor by a peristaltic pump through an L-type joint. Owing to the large ID and simple structure, the injection could be completed within several seconds. The light beam emitted from a halogen lamp was launched into the HF SPR sensor via the MMF. Then the spectrum of the light transmitted through the HF was detected by a spectrometer (HORIBA iHR550). The liquid sensed media used in the experiments are the mixed solutions of polymethylphenyl siloxane fluid and kerosene with different volume ratios. The RIs of these solutions range from 1.51 to 1.58.
4. Results and discussion
Normalized experimental transmission spectra of two HF SPR sensors with different silver layer thicknesses are shown in Fig. 4. The deposition time of silver layer in fabricating is 30s and 40s for Figs. 4(a) and 4(b) respectively. The RIs of the filled liquids (n0) for the curves are 1.5783, 1.5594, 1.5501, 1.5415, 1.5314, 1.5225 and 1.5150 at the wavelength of 589 nm measured by an Abbe refractometer. The RW shifts towards lower wavelength when the RI increases. It is contrary to the solid core optical fiber SPR sensors. With those sensors, RW shifts towards longer wavelength as the RI increases. The HF SPR sensor with a smaller silver layer thickness causes blue shift of the SPR peaks, i.e. for the sensed medium with a particular RI, the RW of the sensor with a thinner silver layer is shorter than that of the sensor with a thicker silver layer. Considering the limited detection range of spectrum, the thicker the silver layer is, the higher RI could be detected. However, when the thickness of silver layer is even larger, the depth of SPR peak decreases gradually and cause detection difficulties.
For the two HFs used in Fig. 4, the thicknesses of the silver layer were theoretically estimated to be approximately 30 nm and 57 nm. The experimental results of RW agree well with theoretically calculation results as shown in Fig. 5(a). The theoretical curves are obtained by the following steps. First calculate the transmission spectra for different n0 with Eq. (5) and find the RW of each spectrum, then plot the theoretical curve of RW with respect to n0. The roughness of the silver layer and the dispersion of the liquid sensed medium were neglected in calculation. The detection range of the sensor depends on the optical spectrum of the light source, the detecting range of the spectrometer, and the silver layer thickness. As the efficient detection range of the CCD detector in the experiment is about from 400 nm to 800 nm, the detection RI range is approximately from 1.509 to 1.763 (the HF with 57 nm silver layer) and 1.494 to 1.697 (the HF with 30 nm silver layer) respectively. As the intensity detected in both ends of spectrum is rather weak, the detection accuracy would decrease when the SPR dip locates near the ends of the spectrum.
The sensitivity of a SPR sensor is defined as ΔRW/Δn0. Here we take the adjacent two points on each curve in Fig. 5(a) as sampling points to calculate the average sensitivities. The HF SPR sensor has a higher sensitivity in the longer wavelength region, i.e. lower RI region of the sensed medium. In other words, the sensitivity scales with the RW. This phenomenon is mainly due to the material dispersion and was extensively elaborated by Homola . The measured sensitivity ranges are 1189~4356 nm/RIU and 2185~6607 nm/RIU for the HFs with 30 nm and 57 nm thick silver layer respectively. These compare favorably with the theoretical values reported for selected coated PCF (maximum of 5500 nm/RIU)  and multi-core holey fiber (2929.39 nm/RIU in the sensing range 1.33-1.42 and 9231.27 nm/RIU in 1.43-1.53) , hollow core optical fiber with a Fiber Bragg Grating (5.93 nm/RIU)  and with long period grating (817 nm/RIU) .
The resolution of a SPR sensor, is typically expressed in terms of the standard deviation of noise of the sensor output,, translated to the refractive index of bulk medium, , where is the bulk refractive index sensitivity [30, 31]. Here, the stability of the spectral system is taken as the standard deviation of the RW calculated from ten repetitive spectral measurements for every n0. The calculated standard deviations are 0.61 nm and 0.55 nm for the 30 nm and 57 nm thick silver layer HF SPR sensors respectively. Thus, with the sensitivity data in the Fig. 5(b), the resolutions are estimated to be 1.4 × 10−4~5.1 × 10−4 RIU and 0.8 × 10−4~2.5 × 10−4 RIU for the HFs with 30 nm and 57 nm thick silver layer respectively as shown in Fig. 5(c). As well as the sensitivity, the sensor with 57 nm thick silver layer has a better resolution. The characteristics of the two HFs are listed in Tab. 1.
Another short length of HF was cut off from each original HF for the observation of scanning electron microscope (SEM). The morphologies of the silver interface and the cross section of the HFs are shown in Fig. 6. The deposition time is 30s for the HF shown in Fig. 6(a) corresponding to the HF used in Fig. 4(a). The deposition time is 40s for the HF in Fig. 6(b) corresponding to the HF in Fig. 4(b). The measured silver layer thicknesses from the SEM pictures are about 30nm and 50nm respectively. The measured thicknesses deviate a little from the calculated results, which is mainly because of the fluctuation in the silver layer thickness shown in the SEM images. Moreover, the silver layer thickness varies slightly along the HF as we adopted liquid phase deposition method. The big bulges in Fig. 6(b) should be fragments or damages of the silver layer generated in the cutting process.
The full width at half maximum (FWHM) of the SPR dip is a parameter that affects the detection accuracy. The broadening of the SPR dip depresses the detection accuracy especially when the signal to noise ratio of the system is low. In Fig. 7(a) we compared the theoretical results with the experimental spectrum when . And the thickness of the silver layer is 57nm. The measured of the output light from the coupling MMF in our experiment set-up is dispersive and decreases linearly with the wavelength (), which is about 10.7° at 400nm and 6.7° at 800nm. The theoretical spectrum using measured dispersive is very close to the measured curve. The calculation results with non-dispersive and smaller angles = 7.5° and = 3.3° are also shown in Fig. 7(a) for comparison, the dashed curves are theoretical results and the solid curve is the measured result. It can be seen that the SPR dip with a smaller is much narrower. It means that a light source with a smaller divergence angle will depress the broadening of the SPR dip and increase detection accuracy. Furthermore, when the in Eq. (2) was a constant, the SPR dip broadens rapidly as it goes into the long wavelength range. However, owing to the gradually decreasing measured of the light source in our experiment, the FWHM of the SPR dip does not increase much along the wavelength. This can be clearly seen in Fig. 7(b) for both calculated and measured data. The theoretical data are got by measuring the FWHM of every theoretical SPR spectrum calculated with Eq. (5) for different n0. The right-most red measured datum is somewhat high. This is mainly due to the extremely low source light power and large in the short wavelength region.
In Fig. 7(b), the theoretical FWHMs of the 30s coating HF are larger than that of the 40s coating HF. The dependence between the theoretical FWHM and the silver thickness of the HF SPR sensor is similar to that of the conventional fiber SPR sensors which has been elaborated by H. Suzuki et al. . The FWHM decreases with respect to the silver thickness, so a smaller FWHM can be obtained with a thicker silver layer. However, the depth of the SPR dip will also decrease as the silver layer goes thicker. Therefore a moderate silver thickness around 50~70 nm is proper. As shown in Fig. 7(b), the measured FWHMs are much larger than the theoretical results for both sensors. This is mainly due to the non-uniformity of the silver layer thickness which can be clearly seen in Fig. 6. The SPR spectrum of a non-uniform silver layer can be simply seen as the combination of the SPR spectra of a set of silver layers with different thicknesses. It will definitely broaden the SPR dip. The larger difference between measured and theoretical FWHMs for the 40s coating HF might be caused by the larger non-uniformity of the silver layer which can also be seen in Fig. 6. Thus contrary to the theoretical results, the 40s coating HF has even larger measured FWHMs than the 30s coating HF. The non-uniformity of the silver layer is hard to avoid for a rather thin silver layer by the chemical silver-plating method, but it can be improved by modifying the silver coating process. In addition, the coupling between the MMF and the HF might also be a factor that increases the FWHM. If the uniformity of the silver layer and the coupling quality are good enough, the FWHM might be close to those solid core fiber SPR sensors under the given experimental set-up.
The figure of merit (FOM), which is defined as the ratio between the sensitivity S and the FWHM, is another characteristic to evaluate the performance of the sensor (FOM = S/FWHM) . Figure 7(c) shows the theoretical and measured FOMs calculated with the data in Fig. 5(b) and Fig. 7(b). The large measured FWHMs of the 40s coating HF cause much lower FOMs than the theoretical results. Although relatively wide FWHMs are obtained in our experiment, the FOMs of the HF SPR sensors are smaller than that of the prism SPR sensors  and close to the conventional solid core fiber SPR sensors .
This study proposed a novel HF SPR sensor applied to the detection of liquids with high RI. The performances of the sensors with different sliver layer thicknesses were investigated both theoretically and experimentally. A measuring system was established and the transmission spectra of the SPR sensor with different RI liquids were measured. The achieved sensitivity is comparable to the sensitivities of other reported fiber SPR sensors. The structure of the designed HF SPR sensor is simple and easy to fabricate. Unlike conventional fiber SPR sensors, the damageable metal layer is intrinsically protected inside the HF, making the sensor more reliable. For the detection of the liquid that might rust the silver layer, a dielectric layer could be coated on the silver layer for protection. If the RI of the coated dielectric layer matches the cladding material, the HF sensor can also support long range SPR. The detection range of the sensor can be extended by changing the RI of cladding layer. The presented technique also provides a new application field in SPR sensing for traditional HF. It can be used in the surface enhancement Raman spectroscopy field if combined with the research of Raman spectroscopy with HF.
This research is supported by the National Natural Science Foundation of China (NSFC) (61201062).
References and links
2. B. Lee, S. Roh, and J. Park, “Current status of micro-and nano-structured optical fiber sensors,” Opt. Fiber Technol. 15(3), 209–221 (2009). [CrossRef]
3. E. Kretchmann and H. Reather, “Radiative decay of non- radiative surface plasmons excited by light,” Z. Naturforsch. A 23, 2135 (1968).
5. K. Kurihara, K. Nakamura, and K. Suzuki, “Asymmetric SPR sensor response curve-fitting equation for the accurate determination of SPR resonance angle,” Sens. Actuators B Chem. 86(1), 49–57 (2002). [CrossRef]
6. W. B. Lin, J. M. Chovelon, and N. Jaffrezic-Renault, “Fiber-optic surface-plasmon resonance for the determination of thickness and optical constants of thin metal films,” Appl. Opt. 39(19), 3261–3265 (2000). [CrossRef] [PubMed]
7. P. Bhatia and B. D. Gupta, “Surface plasmon resonance based fiber optic ammonia sensor utilizing bromocresol purple,” Plasmonics 8(2), 779–784 (2013). [CrossRef]
8. K. Balaa, M. Kanso, S. Cuenot, T. Minea, and G. Louarn, “Experimental realization and numerical simulation of wavelength-modulated fibre optic sensor based on surface plasmon resonance,” Sens. Actuators B Chem. 126(1), 198–203 (2007). [CrossRef]
9. J. F. Masson, Y. C. Kim, L. A. Obando, W. Peng, and K. S. Booksh, “Fiber-optic surface plasmon resonance sensors in the near-infrared spectral region,” Appl. Spectrosc. 60(11), 1241–1246 (2006). [CrossRef] [PubMed]
10. Y. C. Lu, W. P. Huang, and S. S. Jian, “Influence of mode loss on the feasibility of grating-assisted optical fiber surface plasmon resonance refractive index sensors,” J. Lightwave Technol. 27(21), 4804–4808 (2009). [CrossRef]
12. T. Schuster, R. Herschel, N. Neumann, and C. G. Schäffer, “Miniaturized long-period fiber grating assisted surface plasmon resonance sensor,” J. Lightwave Technol. 30(8), 1003–1008 (2012). [CrossRef]
13. Y. Lu, C. J. Hao, B. Q. Wu, M. Musideke, L. C. Duan, W. Q. Wen, and J. Q. Yao, “Surface plasmon resonance sensor based on polymer photonic crystal fibers with metal nanolayers,” Sensors 13(1), 956–965 (2013). [CrossRef] [PubMed]
14. B. Gauvreau, A. Hassani, M. Fassi Fehri, A. Kabashin, and M. A. Skorobogatiy, “Photonic bandgap fiber-based Surface Plasmon Resonance sensors,” Opt. Express 15(18), 11413–11426 (2007). [CrossRef] [PubMed]
15. W. C. Wong, C. C. Chan, J. L. Boo, Z. Y. Teo, Z. Q. Tou, H. B. Yang, C. M. Li, and K. C. Leong, “Photonic Crystal Fiber Surface Plasmon Resonance Biosensor Based on Protein G Immobilization,” IEEE J. Sel. Top. Quant. 19(3), 4602107 (2013). [CrossRef]
16. P. B. Bing, Z. Y. Li, J. Q. Yao, Y. Lu, and Z. G. Di, “A photonic crystal fiber based on surface plasmon resonance temperature sensor with liquid core,” Mod. Phys. Lett. B 26(13), 1250082 (2012). [CrossRef]
17. B. B. Shuai, L. Xia, Y. T. Zhang, and D. M. Liu, “A multi-core holey fiber based plasmonic sensor with large detection range and high linearity,” Opt. Express 20(6), 5974–5986 (2012). [CrossRef] [PubMed]
18. X. Yu, Y. Zhang, S. Pan, P. Shum, M. Yan, Y. Leviatan, and C. Li, “A selectively coated photonic crystal fiber based surface plasmon resonance sensor,” J. Opt. 12(1), 015005 (2010). [CrossRef]
19. A. Hassani and M. Skorobogatiy, “Design of the microstructured optical fiber-based surface plasmon resonance sensors with enhanced microfluidics,” Opt. Express 14(24), 11616–11621 (2006). [CrossRef] [PubMed]
20. G. Nemova and R. Kashyap, “Modeling of plasmon-polariton refractive-index hollow fiber sensors assisted by a fiber Bragg grating,” J. Lightwave Technol. 24(10), 3789–3796 (2006). [CrossRef]
21. L. Xia, Y. Zhang, C. Zhou, B. B. Shuai, and D. M. Liu, “Numerical analysis of plasmon polarition RI fiber sensors with hollow core and a long period grating,” Opt. Commun. 284(12), 2835–2838 (2011). [CrossRef]
24. Y. W. Shi, K. Ito, L. Ma, T. Yoshida, Y. Matsuura, and M. Miyagi, “Fabrication of a polymer-coated silver hollow optical fiber with high performance,” Appl. Opt. 45(26), 6736–6740 (2006). [CrossRef] [PubMed]
25. K. R. Sui, Y. W. Shi, X. L. Tang, X. S. Zhu, K. Iwai, and M. Miyagi, “Optical properties of AgI/Ag infrared hollow fiber in the visible wavelength region,” Opt. Lett. 33(4), 318–320 (2008). [CrossRef] [PubMed]
26. A. K. Sharma and B. D. Gupta, “Theoretical model of a fiber optic remote sensor based on surface plasmon resonance for temperature detection,” Opt. Fiber Technol. 12(1), 87–100 (2006). [CrossRef]
27. Y. Matsuura, M. Saito, M. Miyagi, and A. Hongo, “Loss characteristics of circular hollow waveguides for incoherent infrared light,” J. Opt. Soc. Am. A 6(3), 423 (1989). [CrossRef]
28. M. Born and E. Wolf, Principles of Optics, 7th ed. (Cambridge University, 1999).
29. J. Homola, “On the sensitivity of surface plasmon resonance sensors with spectral interrogation,” Sens. Actuators B Chem. 41(1-3), 207–211 (1997). [CrossRef]
31. H. Y. Lin, C. H. Huang, G. L. Cheng, N. K. Chen, and H. C. Chui, “Tapered optical fiber sensor based on localized surface plasmon resonance,” Opt. Express 20(19), 21693–21701 (2012). [CrossRef] [PubMed]
32. H. Suzuki, M. Sugimoto, Y. Matsui, and J. Kondoh, “Effects of gold film thickness on spectrum profile and sensitivity of a multimode-optical-fiber SPR sensor,” Sens. Actuators B Chem. 132(1), 26–33 (2008). [CrossRef]