An in-line optofluidic refractive index (RI) sensing platform is constructed by splicing a side-channel photonic crystal fiber (SC-PCF) with side-polished single mode fibers. A long-period grating (LPG) combined with an intermodal interference between LP01 and LP11 core modes is used for sensing the RI of the liquid in the side channel. The resonant dip shows a nonlinear wavelength shift with increasing RI over the measured range from 1.3330 to 1.3961. The RI response of this sensing platform for a low RI range of 1.3330-1.3780 is approximately linear, and exhibits a sensitivity of 1145 nm/RIU. Besides, the detection limit of our sensing scheme is improved by around one order of magnitude by introducing the intermodal interference.
© 2016 Optical Society of America
Optofluidics, providing a high-degree integration of easy manipulation of fluids with various kinds of optical mechanisms, has been intensively applied in flow cytometers , liquid dye lasers , label-free biosensors [3,4] and environmental monitoring systems [5–7]. In particular, such integration has greatly contributed to chemical or biological detection and analysis due to its advantages of enabling both accurate control over the flow of liquid samples and real-time optical responses of analyte in liquid [8–12].
Conventional optofluidic sensing platforms consist of microfluidic on-chip waveguides and other off-chip optical components such as light source, lens, and detectors . To comply with the demand of practical, compact and easy handling analytical optofluidic systems, optical fibers are proposed as an alternative choice of on-chip waveguides, providing their distinct features of being free standing and robust, easy light coupling, low guiding losses, low cost for mass produce, and enabling remote sensing. In particular, photonic crystal fiber (PCF) has been regarded as an inherent optofluidic platform due to its unique capability of unprecedented long interaction length between analyte and light by transmitting liquid sample in the micrometer sized channels running along the entire fiber length [13–15]. Moreover, the combination of PCF optofluidic system with measuring refractive index (RI) of the liquid is very promising in analyzing the presence of analyte in the chemical and biological sample, because only small sample volume is needed.
Generally, there are two methods to couple light into PCFs in an optofluidic system, either by splicing PCFs with single mode fibers (SMFs) or applying free-space coupling [16–18]. Splicing with SMFs brings some difficulties in the replacement or manipulation of liquids, while free-space coupling is of less flexibility due to the bulky optical components introduced to the system. To get in-line access to the air holes, drilling holes with femtosecond laser micromachining technique [19,20] and inserting a 20-µm-width C-shaped fiber between the PCF and SMF  have been proposed. However, the micromachining technique requires expensive femtosecond laser and high-accuracy positioning stages. For the latter method, the tiny C-shaped fiber brings difficulty in manipulation, and, more importantly, mode mismatch is unavoidable because C-shaped fiber has weak light guidance capability.
In this paper, we propose and demonstrate an in-line fiber optofluidic RI sensor by splicing a side-channel photonic crystal fiber (SC-PCF) directly to side-polished SMFs. A long-period grating (LPG)  is written on the SC-PCF to provide high sensitivity for monitoring the RI variation of the liquid flow in the side-channel. With the assistance of the LPG and interferometer induced by off-setting splicing, we achieve a broad RI response from 1.3330 to 1.3961 of the liquid circulated in the SC-PCF and an approximately linear response at a low RI range over 1.3330-1.3780 with the sensitivity of 1145 nm/RIU.
2. Configuration and fabrication
The schematic diagram of our design is shown in Fig. 1(a). An SC-PCF with an LPG is spliced to side-polished SMFs at its two ends. The dimensions of the side-polished SMF are specially designed to exactly expose the side channel to enable efficient liquid circulation and light coupling simultaneously. Then, to protect the splicing points, we put each splicing region between a glass plate (indicated by yellow color beneath the fiber in Fig. 1(b)) and a square solidified polydimethylsiloxane (PDMS) chip (indicated by blue color holding the fiber) with a 1.4-mm-diameter hole at the top of it. Extra PDMS is used to seal the space between the fiber and chips to avoid liquid leakage. A syringe pump and a liquid pool are connected to the PDMS chips with polytetrafluoroethylene (PTFE) tubes. Inner diameter and outer diameter of the tube are 1.0 mm and 1.4 mm, respectively.
The SC-PCF used in this experiment is made of pure silica and fabricated through stack-and-draw process . It consists of a solid core surrounded by 6-layer triangular-lattice air holes, as shown as the scanning electron microscope (SEM) image in Fig. 2(a). To enable light-matter interaction and efficient liquid infiltration speed for the application of real-time measurements, a large side channel is created deliberately by removing one sixth of the lattice cladding. The fiber core size and the air filling ratio (the ratio of air hole diameter to the period of the lattice cladding) of the SC-PCF are measured to be 6 µm and 0.6, respectively. The outer cladding diameter is about 118 µm and the diameter of the side channel is measured to be around 30 µm.
Next, a theoretical analysis of the guided modes and their corresponding effective refractive indices (ERIs) is conducted by using finite-element method and importing the real fiber geometry into the simulation model. In order to investigate the guiding properties of the SC-PCF flowing with liquid, a scenario of the side channel infiltrated with water is considered. In the simulations, we also take into consideration of the material dispersion of silica and water according to their Sellmeier equations [23,24]. As the pictures shown in Figs. 2(b) and 2(c), only LP01 and LP11 modes are allowed to propagate in the SC-PCF core within the wavelength range from 1400 nm to 1700 nm when the side channel is filled with water. As a comparison, the guiding properties of the SC-PCF without liquid infiltration are simulated as well. The corresponding guided modes are shown in Figs. 2(d) and 2(e). These two guided modes of the SC-PCF without liquid infiltration are also experimentally confirmed by measuring the mode distributions through a CCD camera, as shown in Figs. 2(f) and 2(g). The LP11 mode profile provided in Fig. 2(g) is obtained by core-offsetting light coupling at the fiber input, and monitoring the mode intensity pattern at the fiber output on a CCD camera. Comparing Figs. 2(b) and 2(c) with Figs. 2(d) and 2(e), more fraction of modal energy is located in the side channel when the side channel is filled with water. It indicates an enhanced light-matter interaction when the side channel is filled with liquid.
The corresponding ERIs of the LP01 and LP11 modes for both conditions are calculated and plotted in Fig. 3(a). As shown, when the medium in the side channel is changed from air to water, the ERIs of LP01 and LP11 core modes increase in a certain degree and the ERI difference between LP01 and LP11 core modes decreases. The pitch of a LPG (), resonant wavelength () and the difference between ERIs of the two coupling modes () satisfy the formula of. To choose appropriate LPG parameters for liquid sensing, we calculate the relationship between the LPG pitch and the resonant wavelength in selective water infiltration condition and plot the results in Fig. 3(b) as the black curve. As a comparison, the relationship between the LPG pitch and the resonant wavelength in air infiltration condition is plotted in Fig. 3(b) as the blue curve. The simulation results indicate that, for our SC-PCF, the LPG resonance shifts to a longer wavelength for the same grating pitch in selective water infiltration condition.
For the fabrication of LPG, the side channel of the SC-PCF is initially infiltrated with water for the purpose of optimizing the wavelength of the resonant dip in the scenario of liquid circulation. To realize the selective infiltration of water into the side channel, firstly, we collapse all the small air holes and leave the side channel open using electric-arc technique  at one fiber end. By dipping this end to liquid container, water is infiltrated into the side channel only through capillaries force. Then, both of the fiber ends are cleaved carefully and spliced to SMFs. The LPG is fabricated through the CO2 laser point-by-point inscription method [26,27]. The number of grating periods is 50. A broadband light source and an optical spectrum analyzer (OSA) are connected to the SC-PCF through SMFs to monitor the spectral variation during the LPG fabrication. We obtain a distinct attenuation band with a central wavelength at around 1520 nm, as shown as the red curve in Fig. 4. The transmission spectra are normalized to the light source spectrum. The actual pitch of the fabricated LPG is measured under a microscope and averaged to be ~153 μm, as shown in the inset of Fig. 3(b). The theoretical grating pitch for resonant wavelength at 1520 nm is 157 μm, as indicated in Fig. 3(b). This slight deviation in the grating pitch may results from 3 reasons. 1) As the desired grating pitch is very close to the focused CO2 beam diameter (in the range of 100 μm to 200 μm), the pitch of the grating fabricated is hard to be precisely controlled. 2) The unavoidable error in the measurement of the grating pitch under microscope is in the order of micrometers. 3) In the experiment, the irradiation of CO2 laser induces a decrease in the ERI difference between LP01 and LP11 modes , and this will result in a blue shift of the resonant wavelength for the same grating pitch. To confirm the mode coupling, we measure the intensity distributions at the fiber output for different wavelengths using a tunable laser source and a camera. As shown as the insets in Fig. 4, the light is efficiently coupled to the LP11 mode near the resonant wavelength (e.g. 1512 nm) and propagates mainly in the LP01 mode at a non-resonant wavelength (e.g. 1570 nm).
To realize an in-line access to the side channel, side-polished SMFs are specially prepared by polishing SMFs up to 36 µm from the center of the fiber core. After aligning carefully the orientation of the SC-PCF and the polished side of the SMFs, we re-splice the SC-PCF to side-polished fibers. In order to achieve high extinction ratio in the resonant band of the LPG, we apply deliberately slight core off-setting splicing between SC-PCF and side-polished SMFs to introduce interference between LP01 and LP11 core modes. To balance the coupling efficiency of the injection light and the interference extinction ratio, the core to core position of the PCF and SMF has an off-setting of around 2 μm. As can be seen from Fig. 4, by introducing the interference, the full-width at half-max (FWHM) value of single resonant dip amplitude is reduced from tens of nanometers to few nanometers.
To analyze the difference between the interference patterns in LPG resonant region and non-resonant region, we theoretically investigate the mechanism of the high-extinction-ratio interference. To simplify the calculation, we assume that the energy transfer of coupling modes occurs in the middle of the LPG . At the incident point, most of the light energy from SMF is coupled to the LP01 core mode in the SC-PCF, and a small fraction to the LP11 mode. After propagating over a length ( = 1.6 cm), the energy of LP01 mode is coupled to LP11 mode at the center of the LPG and continues to pass through the rest of the fiber length ( = 3.6 cm) of SC-PCF as LP11 mode till reaches the convergence splicing point to SMF. Similarly, the energy of weakly excited LP11 mode is coupled to LP01 at the center of LPG, and propagates as LP01 mode in the rest length of SC-PCF, as illustrated in the inset of Fig. 1(c). The two travelling arms interfere at the convergence splicing point to result in the interference pattern. Therefore, the relative phase (Ψ) of the two arms is calculated by the formula as follows :Eq. (3) and the ERIs of LP01 and LP11 modes shown in Fig. 3(a), we can obtain the values of , as plotted in Fig. 3(b) as the red curve. Therefore, the spacing between Dip 2 and Dip 3 (as indicated in Fig. 4) is calculated to be around 15 nm. The slight difference between the calculated spacing and the measured value (11.3 nm) might come from the following reasons. 1) The calculated value of may not be the exactly same as that of the SC-PCF used in the real experiment due to the slight nonuniformity of the fiber introduced in the process of fiber drawing . 2) The measurement error in the fiber length is unavoidable because it is hard to find the precise positions of the starting, middle and ending positions of the LPG fabricated on the SC-PCF. 3) In the theoretical model, we assume that the energy transfer of LP01 and LP11 mode finishes in the middle point of the LPG. In the real experiment, however, the energy transfer of coupling modes may finish before the middle point of the LPG , resulting in an extra deviation between the theoretical resonance spacing and the measured spacing (between Dip 2 and Dip 3).
3. RI sensitivity and discussion
To test the RI sensing performance of this SC-PCF optofluidic sensing platform, we prepare RI solutions with RI ranging from 1.3330 to 1.3961 (measured with Reichert refractometer in the visible wavelength range). During the measurements, the syringe pump is set working in imbibition mode to make the liquid injected from the liquid inlet, flow through the entire length of SC-PCF, and then pulled out from the liquid outlet to the syringe. After one measurement, we pump in a new liquid sample and record a new measurement spectrum. To avoid contamination and ensure the accuracy of experiment results, the solutions are infiltrated in the increasing order of RI. The spectra in Fig. 5(a) show a red shift of the attenuation band when increasing the RI of the liquid.
To analyze the sensing sensitivity, we trace the variation of Dip 2 (indicated in Fig. 5(a)) against the RI change of the liquid in the side channel, and at the same time, run simulations to obtain its theoretical RI sensing sensitivity for comparison. As we take into consideration of the dispersion of water in the simulations, this will result in a different RI value of water (calculated in the near-infrared range) from that we measure with refractometer in the visible range in the experiment . To make the presentation of experimental and simulated data more accurate, we plot the wavelengths of resonance against the relative RI value contrasted with that of water in Fig. 5(b). The wavelength shift of Dip 2 shows a nonlinear relationship against the increase of the liquid RI, and this matches well with the simulation results. This nonlinear response results from the fact that the ERIs of the LP01 and LP11 modes change nonlinearly with the liquid RI in the side channel as indicated in Fig. 3(a). Similar phenomenon has been reported before by V. Bhatia . Besides, within the RI contrast range of 0-0.045, which corresponds to a measured RI range from 1.3330 to 1.3780, the RI response is approximately linear, and the sensitivity over this range is fitted to be around 1145 nm/RIU. Dip 3 shows a slight higher sensitivity of ~1150 nm/RIU. The RI sensitivity of our SC-PCF optofluidic sensor is comparable with that of other highly sensitive fiber optofluidic RI sensor  and LPG based RI sensor . Though some RI sensors [33–35] are reported with sensitivities of one order of magnitude higher than our value, but in their setups, in-line replacement of liquid sample is not available. In our sensing scheme, the realization of in-line liquid manipulation provides much convenience and compactness in real applications.
The detection limit (DL) of a RI sensor is determined by its RI sensitivity () and sensor resolution () by . characterizes the smallest detectable shift of the spectra, taking into consideration of the FWHM value of the resonance (), signal to noise ratio (SNR), spectral resolution and thermal noise. For our device, is dominantly determined by by . As the value of is reduced from 23 nm to 3 nm (Dip 2) at the RI of 1.3330 by introducing the interference, DL of our sensing scheme is improved by approximately 8 times compared with that of the LPG without interferometer. The FWHM value of Dip 3 is smaller (~1.7 nm), giving more than one order of magnitude improvement on the DL value.
We propose and demonstrate an LPG-interferometer assisted SC-PCF in-line optofluidic RI sensing platform for monitoring the liquid circulated in a large side channel. The RI response of a resonant dip has a non-linear increase over a broad RI range of 1.3330-1.3961. The sensitivity of RI response increases with the increasing of liquid RI. In the low RI range of 1.3330 −1.3780 that is common for chemical and biological samples, the RI response is approximately linear with a sensitivity being around 1145 nm/RIU. Moreover, the DL of our sensing scheme is improved by around one order of magnitude by introducing the interferometer. Our SC-PCF based optofluidic RI sensor, which provides the capabilities of in-line manipulation of liquid samples, working in a continuous-flow configuration and detecting RI in a wide range below that of silica with high sensitivity, has great potential for the applications of real-time monitoring and detecting of biological and chemical samples, especially when the volume of analyte is around nano-liters or less.
Singapore Ministry of Education Academic Research Fund Tier 2 (MOE2015-T2-1-066, MOE2014-T2-1-076, MOE2015-T2-2-010); MERLION PhD project from French Ministry of Foreign Affairs; Nanyang Technological University (Startup grant: Lei Wei).
References and links
1. Y. C. Tung, M. Zhang, C. T. Lin, K. Kurabayashi, and S. J. Skerlos, “PDMS-based opto-fluidic micro flow cytometer with two-color, multi-angle fluorescence detection capability using PIN photodiodes,” Sens. Actuators B Chem. 98(2–3), 356–367 (2004). [CrossRef]
2. A. M. Stolyarov, L. Wei, O. Shapira, F. Sorin, S. L. Chua, J. D. Joannopoulos, and Y. Fink, “Microfluidic directional emission control of an azimuthally polarized radial fibre laser,” Nat. Photonics 6(4), 229–233 (2012). [CrossRef]
3. A. A. Yanik, M. Huang, O. Kamohara, A. Artar, T. W. Geisbert, J. H. Connor, and H. Altug, “An optofluidic nanoplasmonic biosensor for direct detection of live viruses from biological media,” Nano Lett. 10(12), 4962–4969 (2010). [CrossRef] [PubMed]
4. Z. He, F. Tian, Y. Zhu, N. Lavlinskaia, and H. Du, “Long-period gratings in photonic crystal fiber as an optofluidic label-free biosensor,” Biosens. Bioelectron. 26(12), 4774–4778 (2011). [CrossRef] [PubMed]
7. K. Li, T. Zhang, G. Liu, N. Zhang, M. Zhang, and L. Wei, “Ultrasensitive optical microfiber coupler based sensors operating near the turning point of effective group index difference,” Appl. Phys. Lett. 109(10), 101101 (2016). [CrossRef]
9. C. Monat, P. Domachuk, and B. J. Eggleton, “Integrated optofluidics: A new river of light,” Nat. Photonics 1(2), 106–114 (2007). [CrossRef]
11. K. Li, G. Liu, Y. Wu, P. Hao, W. Zhou, and Z. Zhang, “Gold nanoparticle amplified optical microfiber evanescent wave absorption biosensor for cancer biomarker detection in serum,” Talanta 120, 419–424 (2014). [CrossRef] [PubMed]
12. Q. Ouyang, S. Zeng, L. Jiang, L. Hong, G. Xu, X.-Q. Dinh, J. Qian, S. He, J. Qu, P. Coquet, and K.-T. Yong, “Sensitivity enhancement of transition metal dichalcogenides/silicon nanostructure-based surface plasmon resonance biosensor,” Sci. Rep. 6, 28190 (2016). [CrossRef] [PubMed]
14. L. Wei, L. Eskildsen, J. Weirich, L. Scolari, T. T. Alkeskjold, and A. Bjarklev, “Continuously tunable all-in-fiber devices based on thermal and electrical control of negative dielectric anisotropy liquid crystal photonic bandgap fibers,” Appl. Opt. 48(3), 497–503 (2009). [CrossRef] [PubMed]
15. H. W. Lee, M. A. Schmidt, P. Uebel, H. Tyagi, N. Y. Joly, M. Scharrer, and P. S. Russell, “Optofluidic refractive-index sensor in step-index fiber with parallel hollow micro-channel,” Opt. Express 19(9), 8200–8207 (2011). [CrossRef] [PubMed]
16. Y. Cui, P. P. Shum, D. J. J. Hu, G. Wang, G. Humbert, and X. Q. Dinh, “Temperature sensor by using selectively filled photonic crystal fiber sagnac interferometer,” IEEE Photonics J. 4(5), 1801–1808 (2012). [CrossRef]
17. N. Zhang, G. Humbert, T. Gong, P. P. Shum, K. Li, J.-L. Auguste, Z. Wu, D. J. J. Hu, F. Luan, Q. X. Dinh, M. Olivo, and L. Wei, “Side-channel photonic crystal fiber for surface enhanced Raman scattering sensing,” Sens. Actuators B Chem. 223, 195–201 (2016). [CrossRef]
18. S. Unterkofler, R. J. McQuitty, T. G. Euser, N. J. Farrer, P. J. Sadler, and P. S. J. Russell, “Microfluidic integration of photonic crystal fibers for online photochemical reaction analysis,” Opt. Lett. 37(11), 1952–1954 (2012). [CrossRef] [PubMed]
20. B. M. Zhang, Y. Lai, W. Yuan, Y. P. Seah, P. P. Shum, X. Yu, and H. Wei, “Laser-assisted lateral optical fiber processing for selective infiltration,” Opt. Express 22(3), 2675–2680 (2014). [CrossRef] [PubMed]
21. C. Wu, M.-L. V. Tse, Z. Liu, B.-O. Guan, C. Lu, and H.-Y. Tam, “In-line microfluidic refractometer based on C-shaped fiber assisted photonic crystal fiber Sagnac interferometer,” Opt. Lett. 38(17), 3283–3286 (2013). [CrossRef] [PubMed]
22. A. M. Vengsarkar, P. J. Lemaire, J. B. Judkins, V. Bhatia, T. Erdogan, and J. E. Sipe, “Long-period fiber gratings as band-rejection filters,” J. Lightwave Technol. 14(1), 58–65 (1996). [CrossRef]
23. I. H. Malitson, “Interspecimen comparison of the refractive index of fused silica,” J. Opt. Soc. Am. 55(10), 1205–1209 (1965). [CrossRef]
24. C. Wu, M.-L. V. Tse, Z. Liu, B.-O. Guan, A. P. Zhang, C. Lu, and H.-Y. Tam, “In-line microfluidic integration of photonic crystal fibres as a highly sensitive refractometer,” Analyst (Lond.) 139(21), 5422–5429 (2014). [CrossRef]
25. G. Humbert and A. Malki, “Characterizations at very high temperature of electric arc-induced long-period fiber gratings,” Opt. Commun. 208(4–6), 329–335 (2002). [CrossRef]
26. Z. Wu, Z. Wang, Y. G. Liu, T. Han, S. Li, and H. Wei, “Mechanism and characteristics of long period fiber gratings in simplified hollow-core photonic crystal fibers,” Opt. Express 19(18), 17344–17349 (2011). [CrossRef] [PubMed]
27. Z. Wu, Y. Liu, Z. Wang, T. Han, S. Li, M. Jiang, P. Ping Shum, and X. Quyen Dinh, “In-line Mach-Zehnder interferometer composed of microtaper and long-period grating in all-solid photonic bandgap fiber,” Appl. Phys. Lett. 101(14), 141106 (2012). [CrossRef]
28. L. Jin, W. Jin, J. Ju, and Y. Wang, “Coupled local-mode theory for strongly modulated long period gratings,” J. Lightwave Technol. 28(12), 1745–1751 (2010). [CrossRef]
29. Z. Sun, Y. G. Liu, Z. Wang, B. Tai, T. Han, B. Liu, W. Cui, H. Wei, and W. Tong, “Long period grating assistant photonic crystal fiber modal interferometer,” Opt. Express 19(14), 12913–12918 (2011). [CrossRef] [PubMed]
34. T. Han, Y. G. Liu, Z. Wang, J. Guo, Z. Wu, S. Wang, Z. Li, and W. Zhou, “Unique characteristics of a selective-filling photonic crystal fiber Sagnac interferometer and its application as high sensitivity sensor,” Opt. Express 21(1), 122–128 (2013). [CrossRef] [PubMed]