Lab on a Chip View Article Online
Published on 08 April 2014. Downloaded by University at Buffalo Libraries on 26/05/2014 15:47:38.
PAPER
Cite this: Lab Chip, 2014, 14, 2090
View Journal | View Issue
An in-fiber integrated optofluidic device based on an optical fiber with an inner core† Xinghua Yang,*a Tingting Yuan,a Pingping Teng,b Depeng Kong,c Chunlan Liu,a Entao Li,a Enming Zhao,a Chengguo Tonga and Libo Yuana A new kind of optofluidic in-fiber integrated device based on a specially designed hollow optical fiber with an inner core is designed. The inlets and outlets are built by etching the surface of the optical fiber without damaging the inner core. A reaction region between the end of the fiber and a solid point obtained after melting is constructed. By injecting samples into the fiber, the liquids can form steady microflows and react in the region. Simultaneously, the emission from the chemiluminescence reaction
Received 12th February 2014, Accepted 7th April 2014
can be detected from the remote end of the optical fiber through evanescent field coupling. The concentration of ascorbic acid (AA or vitamin C, Vc) is determined by the emission intensity of the reaction of Vc, H2O2, luminol, and K3Fe(CN)6 in the optical fiber. A linear sensing range of 0.1–3.0 mmol L−1 for Vc
DOI: 10.1039/c4lc00184b
is obtained. The emission intensity can be determined within 2 s at a total flow rate of 150 μL min−1. Significantly, this work presents information for the in-fiber integrated optofluidic devices without spatial
www.rsc.org/loc
optical coupling.
Introduction Optofluidics, as an emerging field, integrates microfluidics and optics in the same device. Specific advantages such as microfluidic control, a large specific surface area and an integrated light path are obtained.1–6 These intrinsic features of optofluidics allow it to be used in the fields of chemical and environmental analysis, biosynthesis, drug delivery, and other aspects.7–9 Optofluidic devices provide the chance of interaction between materials and light and collect light signals with different kinds of waveguides. Specifically, in these devices, the distribution of the refractive indices of the material would control the propagation path of the light. The refractive index describes the material properties that directly affect how light interacts with different media, such as the degree in which light would be reflected, transmitted, or attenuated. Microstructured optical fibers (MOFs), with microscopic holes penetrating throughout the entire length, have a number of unique characteristics.10–13 The holey structure of MOFs a
Key Laboratory of In-Fiber Integrated Optics, Ministry of Education, College of Science, Harbin Engineering University, Harbin 150001, China. E-mail: [email protected] b Academy of Harbin North Special Vehicle, Harbin First Machinery Manufacturing Group Co. Ltd, Harbin 150056, China c State Key Laboratory of Transient Optics and Photonics, Xi'an Institute of Optics and Precision Mechanic, Chinese Academy of Sciences, Xi'an 710119, China † Electronic supplementary information (ESI) available. See DOI: 10.1039/ c4lc00184b
2090 | Lab Chip, 2014, 14, 2090–2095
has an inherent advantage of being able to hold small volumes of fluid. Such a small volume is very useful for microanalysis and can simplify the experimental setup of optical fiber sensors.14–16 Especially, a fraction of the modal field of solid-core MOFs is located within the air-holes of the cladding. This allows the direct interaction of the guided light with different gases, liquids or biological samples via evanescent field effects within the whole length of the fiber. Their one-dimension holey or tubular microstructure greatly enhances the specific surface area for sensing and prevents the sensing layers from being damaged compared to traditional optical fibers. These advantages have attracted much attention in fabricating optical fiber devices. Further, the low consumption of reagents and the high specific surface area in the sensing devices based on these optical fibers result in better mass transfer and a fast sensing process than macrosystems of analysis.17 Recently, great efforts have been made to investigate MOF-based chemical sensors in our previous work. On the other hand, as an advanced technique of analysis, chemiluminescence (CL) has received remarkable attention.18–25 It is characterized by high sensitivity and simple instrumentation compared with other spectrophotometric techniques. In comparison with other optical detection methods such as fluorescence, Raman spectroscopy and absorption, the process of CL has a wide linear range, uses an inexpensive apparatus, has a fast response time, high sensitivity, and no interference from background scattering light. It has been widely applied to determine inorganic and
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 08 April 2014. Downloaded by University at Buffalo Libraries on 26/05/2014 15:47:38.
Lab on a Chip
organic species such as drugs, pollutants and pesticides in the fields of clinic, environment, agriculture and industry. In this paper, we present a new kind of in-fiber integrated chemiluminescence optofluidic device based on a specially designed optical fiber with an inner core. This optofluidic device can not only enable an intense and uniform light distribution by improving the photon transport but also realize adequate light coupling between fluid and waveguides by enhancing the mass transport and the evanescent field along the core of the optical fiber. Specifically, in the design, the path for the microfluid of the CL reagents is built by etching microholes on the surface of this specially designed optical fiber without damaging the structure of the core. In comparison with previous devices based on hollow-core fiber techniques, the inlet and the outlet of the microfluids on the surface of this optical fiber are very suitable for on-line sampling. Here, we choose one compound, ascorbic acid (AA) or vitamin C (Vc), to demonstrate the working principle of this device. Vc is a powerful antioxidant which is present in food and beverages and is often used as a chemical marker for evaluating food deterioration, product quality, and freshness.26 Furthermore, Vc can help to promote healthy cell development, iron absorption, and normal tissue growth, playing significant roles in the proper functioning of human metabolism and central nervous and renal systems.27–31 In this design, the reaction of CL occurs in the optical fiber and the emitted light can be coupled into the fiber through the evanescent wave field, and then the concentration of Vc can be determined by the intensity of the CL from the end of the optical fiber.
Experimental Design and fabrication of the optical fiber with an inner core To construct the in-fiber microfluidic device, a special hollow optical fiber with an inner core based on silica was designed and fabricated in the lab. The structure of this special optical fiber can simultaneously provide a microfluidic cave and the core of the waveguide during testing. This compatibility is hard to find in traditional optical fibers and other MOFs. Firstly, a preform was fabricated using an assembling method following the usual way in our lab.32–34 Secondly, the prepared preform was then drawn into the optical fiber. Many parameters such as the heating temperature, translation speed, drawing speed and pressure in the hole were controlled during this process. By selecting suitable sets of parameters, we can completely control the profiles of the fibers. By adjusting the parameters, fibers with hole diameters of around 20–180 μm and core diameters varied within the range of 10 to 40 μm can be obtained. The structure and the refractive index distribution of the optical fiber with a length of 15 cm are described in Fig. 1(a). The fiber has a hollow structure and an inner core in it. The outer diameter of the fiber is about 300 μm, the inner diameter of the fiber is about 180 μm and the diameter of the core is about 30 μm. This optical fiber was simulated using the beam propagation
This journal is © The Royal Society of Chemistry 2014
Paper
Fig. 1 (a) The distribution of the refractive index of the optical fiber with an inner core. Inset: the cross section of the optical fiber. (b) The simulation result of the optical fiber using the beam propagation method (BPM). The refractive index of the core is higher than that of the circular cladding so that the light can be guided in the core.
method (BPM) software to properly describe the optical confinement loss of the waveguide structure.35,36 Fig. 1(b) illustrates the simulation result of the optical fiber when the hollow cave is filled with water solution (n = 1.33). In this simulation, the refractive index of the circular cladding is 1.45 and that of the core is 1.46. When the central cave is filled with the microfluidic solution, the light is guided in
Fig. 2 The structural principle of the optofluidic device. Inset: the morphologies of the microholes and the solid point obtained after melting the optical fiber with an inner core. From the cross section of the optical fiber at the melting region when guiding blue light, we can observe that the whole optical fiber collapses to a solid waveguide at the melting point. Because the refractive index of the core is higher than that of the cladding, the light can still be guided along the core at this region.
Lab Chip, 2014, 14, 2090–2095 | 2091
View Article Online
Paper
Lab on a Chip
the core with a higher refractive index and more than 90% of the total guided light power is confined within the core.
Published on 08 April 2014. Downloaded by University at Buffalo Libraries on 26/05/2014 15:47:38.
Principle and experimental setup Based on this specially designed hollow structure with an inner core, both the on-line injection of the sample and in-fiber CL reaction were realized by side-opening microholes on the surface of the optical fiber (Fig. 2). Specifically, the detection process involved two solutions, luminol (5-amino2,3-dihydro-1,4-phthalazinedione) solution and a mixture of potassium ferricyanide (K3Fe(CN)6), H2O2 and ascorbic acid samples. In this design, one end of the optical fiber was used as the inlet of the samples. Moreover, to realize the injection of the light-emitting reagent of luminol and ensure that CL takes place inside the fiber, one microhole of the inlet was opened on the surface of the optical fiber by means of CO2 laser etching. Here, we collimated a beam of a CO2 laser at λ = 10.6 μm using a Keplerian telescopic system on the surface of the optical fiber.37 The laser with a Gaussian beam profile was focused on the exposed fiber and was computer programmed to scan across the fiber point-wise in the transverse direction. The CO2 laser power was set at 8 W and the speed of scanning was 100 mm s−1. The frequency of the laser was 20 kHz. The microholes were obtained by repeatedly scanning 120 times. The distance between the left end of the fiber and the left hole was 1 cm. The distance between the two holes was 10 cm. To reduce the dead volume, the distance between the right hole and the melted point was only 1 mm. The whole process of micromachining was monitored using a microscope system. In addition, the diameter of the fiber is of the order of the laser beam pot size and when the fiber is placed a little outside the Rayleigh range of the beam,38 the beam spot size is larger and usually the etching range is wider. Then, by adjusting the position of the fiber, the diameter of the hole can be controlled. In this work, the spot size for the Gaussian beam is 50 μm and we placed the optical fiber at the focus of the beam. As shown by the inset of the figure, the width of the holes is about 50 μm and the morphology is uniform without obvious melting effect beside the etching point. During the process of etching, the inner core is placed away from the processing point so that the structure of the waveguide is not damaged. After detection, the waste solution can outflow through another opened microhole of the outlet after CL reaction. Additionally, to prevent the waste solution from moving towards the remote opening end of the optical fiber, we cut off the microflow of the liquid by melting the core and the cladding to form a solid point beside the microhole of the outlet using a welding machine (see the right figure in the inset of Fig. 2). This will also avoid the influence of the liquid at the end on the optical coupling between the optical fiber and the detector. The configuration of the in-fiber integrated optofluidic device is shown in Fig. 3. The optical fiber was placed on the surface of a silica substrate. The two solutions were separately
2092 | Lab Chip, 2014, 14, 2090–2095
Fig. 3 A sketch of the in-fiber integrated CL optofluidic device. Inset: the structure of the connector.
contained with two injection pumps and connected with the connectors using polytetrafluoroethene (PTFE) capillaries (the outer diameter is 1.5 mm and the inner diameter is 0.5 mm). The PTFE capillaries were inserted into the connectors at the left end, the left inlet and outlet. They were immobilized using screws. The connectors were immobilized on the silica substrate using an adhesive. The solution can be injected into the optical fiber through the microholes without leakage. The inset of Fig. 3 presents the sketch of their connectivity to the optical fiber. At the beginning of CL reactions, light is emitted from the reagents and guided by the core of the optical fiber through evanescent field coupling39,40 of the core. The optical fiber provides enough length of the core for the collection of the light. Simultaneously, the light in the core can be detected at the remote end using a photomultiplier tube (PMT, R2949; Hamamatsu, Japan). The PMT is equipped with a high accuracy negative power source (biased at −1200 V for measurement) independently and connected with a photon counter. To avoid the interference of external light, the region within the red dashed box in Fig. 3 is placed in a dark box.
Results and discussion In this in-fiber integrated optofluidic CL analyzing system, two streams of fluids were pumped into the optical fiber. The first is a mixture of Vc, H2O2 and K3Fe(CN)6 and the other is luminol solution. K3Fe(CN)6 standard solution (1.0 × 10−2 M) was prepared by dissolving 0.330 g in a volumetric flask of 100 mL with distilled water. Vc standard solution was prepared by dissolving 3.649 g in a volumetric flask of 100 mL with distilled water. This solution was diluted to a specific concentration before use. H2O2 was used as the oxidant and prepared by diluting a commercial H2O2 solution with a concentration of 30% (w/w) to 0.4 mol L−1. Before injection, Vc, K3Fe(CN)6 and H2O2 solutions were mixed with the same volume. Luminol standard solution (1.0 × 10−2 M) was prepared by dissolving 0.272 g of luminol (Sigma-Aldrich) in a volumetric flask of 150 mL with NaOH solution (pH = 11). All of the
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 08 April 2014. Downloaded by University at Buffalo Libraries on 26/05/2014 15:47:38.
Lab on a Chip
standard solutions were stored at 4 °C in a refrigerator before use. When the above solutions were injected into the optical fiber, all of the streams were driven by syringe pumps at a fixed flow rate of 150 μl min−1 (the rate of each syringe pump is 75 μL min−1) which is an optimum value as discussed later. The typical spectrum of the CL reaction of luminol, K3Fe(CN)6, and H2O2 was obtained using a spectrometer (Fig. 4). From this curve, we can observe that this reaction has a typical emission of blue light and the maximum CL intensity is at around 430 nm. In this reaction, luminol is oxidized by H2O2 under the catalysis of K3Fe(CN)6. During the reaction, OH− removes the nitrogen protons, releasing a negative charge, which moves onto the carbonyl oxygen to form an enolate. Then, the oxidant (decomposed from H2O2 under the catalysis of K3Fe(CN)6) performs a cycloaddition reaction with the carbonyl carbons, forming an excited state (see the inset in Fig. 2) which emits blue light. Then, the photon is efficiently coupled into the inner core through the evanescent field and can be detected using the PMT. In this process, Vc influences the above reaction and the CL intensity depends on the concentration of Vc. To enhance the emission, the CL reaction was carried out in alkaline solution. The luminol solution was prepared in NaOH alkaline solutions with different pH values. Results indicate that CL reaction can produce the strongest emission in NaOH solution with a pH value of 11.0. Hence, this alkaline solution was chosen for the in-fiber integrated optofluidic analysis of Vc. When the two solutions were pumped hydrodynamically into the end and the opened inlet, the signal was remotely detected downstream of the point-of-confluence. The CL signal was observed to rise monotonically from zero to the maximum steady-state value within one second when the reagents were mixed. To present the change of the light intensity in the inner core when adjusting the concentration of Vc in the optofluidic reaction, the core at the end of the optical fiber was observed using a microscope. Fig. 5(a) shows that when the concentration of Vc increased from 0 to 2.81 mmol L−1, the intensity of the blue light at the end greatly decreased. The relationship between the CL intensity and the concentrations of Vc is presented in Fig. 5(b). The normalized CL
Fig. 4 The emission spectrum of the reaction of CL.
This journal is © The Royal Society of Chemistry 2014
Paper
Fig. 5 (a) Changes of the light intensity in the inner core during the optofluidic reaction. The concentration variation of Vc in the optical fiber: top left: 0 mmol L−1, top right: 0.65 mmol L−1, bottom left: 2.2 mmol L−1 and bottom right: 2.81 mmol L−1. (b) The relationship between the CL intensity and the concentration of Vc. Inset (top): the typical calibration plots for the Vc samples in the range of 0–3.0 mmol L−1. Inset (bottom): CL emission in the lateral view of the special optical fiber observed with a microscope.
intensity detected from the optical fiber with the air core is plotted against the concentration of Vc. From the response curve, we found that the intensity is obviously weakened when the concentration of Vc increases. The CL response of the device is nearly linear depending on the concentration of Vc in the range of about 0 to 3.0 mmol L−1 (inset (top) in Fig. 5(b)). In this range, the highest concentration of Vc is mainly determined by the lowest CL intensity which can be detected. Additionally, the inset (bottom) of Fig. 5(b) shows the CL emission in the lateral view of the special optical fiber observed with a microscope. Using the microscope, we can visually observe the emission change of the CL reaction with different concentrations of the samples. In this design, the intensity of the CL signal depends on the mixing dynamics within the optical fiber. At the optimal flow rate, reagents can be fully mixed by the time they reach the outlet, leading to a strong CL signal. To determine the optimal flow conditions for the experiments, we fixed the concentration of Vc at 0.2 mmol L−1 and investigated the influence of the total flow rate on the measured CL signal. The variation of the CL signal in response to changes of the flow rate is shown in Fig. 6. At slow flow rates such as 100 μl min−1,
Lab Chip, 2014, 14, 2090–2095 | 2093
View Article Online
Paper
Lab on a Chip
involving the process of CL or fluorescence quenching can be further integrated with this structure of the optical fiber.
Published on 08 April 2014. Downloaded by University at Buffalo Libraries on 26/05/2014 15:47:38.
Acknowledgements This work is supported by the National Natural Science Foundation of China (NSFC, 61007053, 61205027, 61377085, 61177081, 61227013), the 111 project (B13015) to the Harbin Engineering University, and the Fundamental Research Funds for the Central Universities.
Notes and references Fig. 6 Dynamic response of the in-fiber analyzing system.
the reaction is partially over within a short length of the optical fiber, leading again to a reduced signal. At higher flow rates such as 200 μl min−1, the reaction is inadequate by the time they reach the outlet and leads to a reduced CL signal. As a result, the strongest CL signal was obtained at an appropriate total flow rate of 150 μl min−1. In practice, the total flow rate was fixed at 150 μL min−1 to investigate the influence of Vc concentration on the intensity of the CL signal. Additionally, the dynamic response in Fig. 6 indicates that the steady-state CL emission can be obtained within 2 s. Conversely, when the sample injection is stopped, the emission quickly decreases and can also reach the minimum value within this time. These results also reveal that the sample solutions can form steady microflows around the core in the cave of the optical fiber.
Conclusions We describe a kind of optofluidic in-fiber integrated device based on a specially designed hollow optical fiber with an inner core. Especially, the path of the microfluid was built by etching microholes on the surface of this specially designed optical fiber without damaging the structure of the core which can be hardly realized in other kinds of MOFs. The results show that the CL reagents and samples can be mixed and form steady microflows in the optical fiber. Simultaneously, the emission from the reaction is efficiently coupled into the inner core through the long distance evanescent field, guided along the core through the melted point and then detected from the remote end of the optical fiber. Results indicate that the in-fiber integrated optofluidic device can achieve an efficient optical signal from the inner core while reducing the volume of the device. By analyzing the emission intensity of the fiber core, the linear sensing range of 0.1–3.0 mmol L−1 for Vc is obtained. The dynamic response reveals that the emission intensity can be determined within 2 s at a total flow rate of 150 μL min−1 (the rates of the two injection pumps are equal to 75 μL min−1). In addition, other in-fiber integrated optofluidic devices analyzing flowable substances such as some liquids and gases
2094 | Lab Chip, 2014, 14, 2090–2095
1 D. Psaltis, S. Quake and C. Yang, Nature, 2006, 442, 381–386. 2 X. D. Fan and S. Yun, Nat. Methods, 2014, 11, 141–147. 3 P. Müller, D. Kopp, A. Llobera and H. Zappe, Lab Chip, 2014, 14, 737–743. 4 I. Rodríguez-Ruiz, A. Llobera, J. Vila-Planas, D. Johnson, J. Gómez-Morales and J. García-Ruiz, Anal. Chem., 2013, 85, 9678–9685. 5 A. Mitra, B. Deutsch, F. Ignatovich, C. Dykes and L. Novotny, ACS Nano, 2010, 4, 1305–1312. 6 Z. Meng, X. Zhang and J. Qin, Nanoscale, 2013, 5, 4687–4690. 7 H. C. Fan, J. Wang, A. Potanina and S. R. Quake, Nat. Biotechnol., 2011, 29, 51–57. 8 K. H. Rasmussen, R. Marie, J. M. Lange, W. E. Svendsen, A. Kristensena and K. U. Mirb, Lab Chip, 2011, 11, 1431–1433. 9 J. Y. Liu and Y. X. Duan, Oral Oncol., 2012, 48, 569–577. 10 R. He, P. J. A. Sazio, A. C. Peacock, N. Healy, J. R. Sparks, M. Krishnamurthi, V. Gopalan and J. V. Badding, Nat. Photonics, 2012, 6, 174–179. 11 Y. Wang, D. Wang, C. Liao, T. Hu, J. Guo and H. Wei, Opt. Lett., 2013, 38, 269–271. 12 W. Qian, C. Zhao, Y. Wang, C. Chan, S. Liu and W. Jin, Opt. Lett., 2010, 34, 3296–3298. 13 X. H. Yang and L. Wang, Opt. Express, 2007, 15, 16478–16483. 14 Y. L. Ruan, H. Ebendorff-Heidepriem, S. Afshar and T. M. Monro, Opt. Express, 2010, 25, 26018–26026. 15 J. J. Kaufman, G. M. Tao, S. Shabahang, E. H. Banaei, D. S. Deng, X. Liang, S. G. Johnson, Y. Fink and A. F. Abouraddy, Nature, 2012, 487, 463–467. 16 Y. P. Wang, X. L. Tan, W. Jin, D. Q. Ying, Y. L. Hoo and S. J. Liu, Opt. Lett., 2010, 35, 88–90. 17 X. Yu, Y. C. Kwok, N. A. Khairudin and P. Shum, Sens. Actuators, B, 2009, 137, 462–466. 18 I. Al-Ogaidi, H. Gou, Z. Aguilar, S. Guo, A. Melconian, A. Al-kazaz, F. Mengand and N. Wu, Chem. Commun., 2014, 50, 1344–1346. 19 S. Casolari, B. Roda, M. Mirasoli, M. Zangheri, D. Patrono, P. Reschiglian and A. Roda, Analyst, 2013, 138, 211–219. 20 M. Kaczmarek and S. Lis, Anal. Methods, 2012, 4, 1964–1967. 21 B. Rezaei, A. Ensafi and F. Haghighatnia, Anal. Methods, 2012, 4, 1573–1578. 22 A. Natrajan, D. Sharpe and D. Wen, Org. Biomol. Chem., 2012, 10, 3432–3447.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 08 April 2014. Downloaded by University at Buffalo Libraries on 26/05/2014 15:47:38.
Lab on a Chip
23 K. Lok, S. Muttalib, P. Lee, Y. Kwok and N. Nguyen, Lab Chip, 2012, 12, 2353–2361. 24 D. J. Bowles, A. S. Olson, N. Ii, D. L. Heglund, R. L. Vargas and S. A. Williams, Appl. Spectrosc., 2012, 66, 175–17. 25 A. Stolyarov, A. Gumennik, W. McDaniel, O. Shapira, B. Schell, F. Sorin, K. Kuriki, G. Benoit, A. Rose, J. Joannopoulos and Y. Fink, Opt. Express, 2012, 20, 12407–12415. 26 N. Moghimi and K. T. Leung, Anal. Chem., 2013, 85, 5974–5980. 27 T. Huang, L. Yeh, H. Huang, Z. Nian, Y. Yeh, Y. Chou, J. Yeh and M. Tsai, Polym. Chem., 2014, 5, 630–637. 28 C. Sun, C. Chang, H. Lee, J. Zhou, J. Wang, T. Sham and W. Pong, ACS Nano, 2011, 5, 7788–7795. 29 L. Li, X. Cai, Y. Ding, S. Gu and Q. Zhang, Anal. Methods, 2013, 5, 6748–6754. 30 C. Weng, P. Hsu, S. Hsu, C. Chang, W. Hung, P. Wu and J. Yeh, J. Mater. Chem. B, 2013, 1, 4983–4991. 31 Q. Ma, Y. Li, Z. Lin, G. Tang and X. Su, Nanoscale, 2013, 5, 9726–9731.
This journal is © The Royal Society of Chemistry 2014
Paper
32 X. H. Yang, Y. Zheng, S. Luo, Y. Liu and L. B. Yuan, Sens. Actuators, B, 2013, 182, 571–575. 33 X. H. Yang, Y. Liu, Y. Zheng, S. Li, L. B. Yuan, T. Yuan and C. Tong, Opt. Commun., 2013, 304, 83–86. 34 A. Zhou, Y. Zhang, G. Li, J. Yang, Y. Wang, F. Tian and L. B. Yuan, Opt. Lett., 2011, 36, 3221–3223. 35 N. Mothe and P. Bin, Opt. Express, 2009, 17, 15778–15789. 36 J. Kim, J. Kim, Y. M. Jung, W. Ha, Y. Jeong, S. Lee, A. Tünnermann and K. Oh, Opt. Lett., 2009, 34, 2973–2975. 37 Y. J. Rao, Y. P. Wang, Z. L. Ran and T. Zhu, J. Lightwave Technol., 2003, 21, 1320–1327. 38 L. Lévesque and V. Jdanov, Opt. Laser Technol., 2010, 42, 1080–1083. 39 P. Lin, S. W. Kwok, H. Lin, V. Singh, L. Kimerling, G. Whitesides and A. Agarwal, Nano Lett., 2014, 14, 231–238. 40 Á. López-Lorente, M. Sieger, M. Valcárcel and B. Mizaikoff, Anal. Chem., 2014, 86, 783–789.
Lab Chip, 2014, 14, 2090–2095 | 2095
Published on 08 April 2014. Downloaded by University at Buffalo Libraries on 26/05/2014 15:47:38.
PAPER
Cite this: Lab Chip, 2014, 14, 2090
View Journal | View Issue
An in-fiber integrated optofluidic device based on an optical fiber with an inner core† Xinghua Yang,*a Tingting Yuan,a Pingping Teng,b Depeng Kong,c Chunlan Liu,a Entao Li,a Enming Zhao,a Chengguo Tonga and Libo Yuana A new kind of optofluidic in-fiber integrated device based on a specially designed hollow optical fiber with an inner core is designed. The inlets and outlets are built by etching the surface of the optical fiber without damaging the inner core. A reaction region between the end of the fiber and a solid point obtained after melting is constructed. By injecting samples into the fiber, the liquids can form steady microflows and react in the region. Simultaneously, the emission from the chemiluminescence reaction
Received 12th February 2014, Accepted 7th April 2014
can be detected from the remote end of the optical fiber through evanescent field coupling. The concentration of ascorbic acid (AA or vitamin C, Vc) is determined by the emission intensity of the reaction of Vc, H2O2, luminol, and K3Fe(CN)6 in the optical fiber. A linear sensing range of 0.1–3.0 mmol L−1 for Vc
DOI: 10.1039/c4lc00184b
is obtained. The emission intensity can be determined within 2 s at a total flow rate of 150 μL min−1. Significantly, this work presents information for the in-fiber integrated optofluidic devices without spatial
www.rsc.org/loc
optical coupling.
Introduction Optofluidics, as an emerging field, integrates microfluidics and optics in the same device. Specific advantages such as microfluidic control, a large specific surface area and an integrated light path are obtained.1–6 These intrinsic features of optofluidics allow it to be used in the fields of chemical and environmental analysis, biosynthesis, drug delivery, and other aspects.7–9 Optofluidic devices provide the chance of interaction between materials and light and collect light signals with different kinds of waveguides. Specifically, in these devices, the distribution of the refractive indices of the material would control the propagation path of the light. The refractive index describes the material properties that directly affect how light interacts with different media, such as the degree in which light would be reflected, transmitted, or attenuated. Microstructured optical fibers (MOFs), with microscopic holes penetrating throughout the entire length, have a number of unique characteristics.10–13 The holey structure of MOFs a
Key Laboratory of In-Fiber Integrated Optics, Ministry of Education, College of Science, Harbin Engineering University, Harbin 150001, China. E-mail: [email protected] b Academy of Harbin North Special Vehicle, Harbin First Machinery Manufacturing Group Co. Ltd, Harbin 150056, China c State Key Laboratory of Transient Optics and Photonics, Xi'an Institute of Optics and Precision Mechanic, Chinese Academy of Sciences, Xi'an 710119, China † Electronic supplementary information (ESI) available. See DOI: 10.1039/ c4lc00184b
2090 | Lab Chip, 2014, 14, 2090–2095
has an inherent advantage of being able to hold small volumes of fluid. Such a small volume is very useful for microanalysis and can simplify the experimental setup of optical fiber sensors.14–16 Especially, a fraction of the modal field of solid-core MOFs is located within the air-holes of the cladding. This allows the direct interaction of the guided light with different gases, liquids or biological samples via evanescent field effects within the whole length of the fiber. Their one-dimension holey or tubular microstructure greatly enhances the specific surface area for sensing and prevents the sensing layers from being damaged compared to traditional optical fibers. These advantages have attracted much attention in fabricating optical fiber devices. Further, the low consumption of reagents and the high specific surface area in the sensing devices based on these optical fibers result in better mass transfer and a fast sensing process than macrosystems of analysis.17 Recently, great efforts have been made to investigate MOF-based chemical sensors in our previous work. On the other hand, as an advanced technique of analysis, chemiluminescence (CL) has received remarkable attention.18–25 It is characterized by high sensitivity and simple instrumentation compared with other spectrophotometric techniques. In comparison with other optical detection methods such as fluorescence, Raman spectroscopy and absorption, the process of CL has a wide linear range, uses an inexpensive apparatus, has a fast response time, high sensitivity, and no interference from background scattering light. It has been widely applied to determine inorganic and
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 08 April 2014. Downloaded by University at Buffalo Libraries on 26/05/2014 15:47:38.
Lab on a Chip
organic species such as drugs, pollutants and pesticides in the fields of clinic, environment, agriculture and industry. In this paper, we present a new kind of in-fiber integrated chemiluminescence optofluidic device based on a specially designed optical fiber with an inner core. This optofluidic device can not only enable an intense and uniform light distribution by improving the photon transport but also realize adequate light coupling between fluid and waveguides by enhancing the mass transport and the evanescent field along the core of the optical fiber. Specifically, in the design, the path for the microfluid of the CL reagents is built by etching microholes on the surface of this specially designed optical fiber without damaging the structure of the core. In comparison with previous devices based on hollow-core fiber techniques, the inlet and the outlet of the microfluids on the surface of this optical fiber are very suitable for on-line sampling. Here, we choose one compound, ascorbic acid (AA) or vitamin C (Vc), to demonstrate the working principle of this device. Vc is a powerful antioxidant which is present in food and beverages and is often used as a chemical marker for evaluating food deterioration, product quality, and freshness.26 Furthermore, Vc can help to promote healthy cell development, iron absorption, and normal tissue growth, playing significant roles in the proper functioning of human metabolism and central nervous and renal systems.27–31 In this design, the reaction of CL occurs in the optical fiber and the emitted light can be coupled into the fiber through the evanescent wave field, and then the concentration of Vc can be determined by the intensity of the CL from the end of the optical fiber.
Experimental Design and fabrication of the optical fiber with an inner core To construct the in-fiber microfluidic device, a special hollow optical fiber with an inner core based on silica was designed and fabricated in the lab. The structure of this special optical fiber can simultaneously provide a microfluidic cave and the core of the waveguide during testing. This compatibility is hard to find in traditional optical fibers and other MOFs. Firstly, a preform was fabricated using an assembling method following the usual way in our lab.32–34 Secondly, the prepared preform was then drawn into the optical fiber. Many parameters such as the heating temperature, translation speed, drawing speed and pressure in the hole were controlled during this process. By selecting suitable sets of parameters, we can completely control the profiles of the fibers. By adjusting the parameters, fibers with hole diameters of around 20–180 μm and core diameters varied within the range of 10 to 40 μm can be obtained. The structure and the refractive index distribution of the optical fiber with a length of 15 cm are described in Fig. 1(a). The fiber has a hollow structure and an inner core in it. The outer diameter of the fiber is about 300 μm, the inner diameter of the fiber is about 180 μm and the diameter of the core is about 30 μm. This optical fiber was simulated using the beam propagation
This journal is © The Royal Society of Chemistry 2014
Paper
Fig. 1 (a) The distribution of the refractive index of the optical fiber with an inner core. Inset: the cross section of the optical fiber. (b) The simulation result of the optical fiber using the beam propagation method (BPM). The refractive index of the core is higher than that of the circular cladding so that the light can be guided in the core.
method (BPM) software to properly describe the optical confinement loss of the waveguide structure.35,36 Fig. 1(b) illustrates the simulation result of the optical fiber when the hollow cave is filled with water solution (n = 1.33). In this simulation, the refractive index of the circular cladding is 1.45 and that of the core is 1.46. When the central cave is filled with the microfluidic solution, the light is guided in
Fig. 2 The structural principle of the optofluidic device. Inset: the morphologies of the microholes and the solid point obtained after melting the optical fiber with an inner core. From the cross section of the optical fiber at the melting region when guiding blue light, we can observe that the whole optical fiber collapses to a solid waveguide at the melting point. Because the refractive index of the core is higher than that of the cladding, the light can still be guided along the core at this region.
Lab Chip, 2014, 14, 2090–2095 | 2091
View Article Online
Paper
Lab on a Chip
the core with a higher refractive index and more than 90% of the total guided light power is confined within the core.
Published on 08 April 2014. Downloaded by University at Buffalo Libraries on 26/05/2014 15:47:38.
Principle and experimental setup Based on this specially designed hollow structure with an inner core, both the on-line injection of the sample and in-fiber CL reaction were realized by side-opening microholes on the surface of the optical fiber (Fig. 2). Specifically, the detection process involved two solutions, luminol (5-amino2,3-dihydro-1,4-phthalazinedione) solution and a mixture of potassium ferricyanide (K3Fe(CN)6), H2O2 and ascorbic acid samples. In this design, one end of the optical fiber was used as the inlet of the samples. Moreover, to realize the injection of the light-emitting reagent of luminol and ensure that CL takes place inside the fiber, one microhole of the inlet was opened on the surface of the optical fiber by means of CO2 laser etching. Here, we collimated a beam of a CO2 laser at λ = 10.6 μm using a Keplerian telescopic system on the surface of the optical fiber.37 The laser with a Gaussian beam profile was focused on the exposed fiber and was computer programmed to scan across the fiber point-wise in the transverse direction. The CO2 laser power was set at 8 W and the speed of scanning was 100 mm s−1. The frequency of the laser was 20 kHz. The microholes were obtained by repeatedly scanning 120 times. The distance between the left end of the fiber and the left hole was 1 cm. The distance between the two holes was 10 cm. To reduce the dead volume, the distance between the right hole and the melted point was only 1 mm. The whole process of micromachining was monitored using a microscope system. In addition, the diameter of the fiber is of the order of the laser beam pot size and when the fiber is placed a little outside the Rayleigh range of the beam,38 the beam spot size is larger and usually the etching range is wider. Then, by adjusting the position of the fiber, the diameter of the hole can be controlled. In this work, the spot size for the Gaussian beam is 50 μm and we placed the optical fiber at the focus of the beam. As shown by the inset of the figure, the width of the holes is about 50 μm and the morphology is uniform without obvious melting effect beside the etching point. During the process of etching, the inner core is placed away from the processing point so that the structure of the waveguide is not damaged. After detection, the waste solution can outflow through another opened microhole of the outlet after CL reaction. Additionally, to prevent the waste solution from moving towards the remote opening end of the optical fiber, we cut off the microflow of the liquid by melting the core and the cladding to form a solid point beside the microhole of the outlet using a welding machine (see the right figure in the inset of Fig. 2). This will also avoid the influence of the liquid at the end on the optical coupling between the optical fiber and the detector. The configuration of the in-fiber integrated optofluidic device is shown in Fig. 3. The optical fiber was placed on the surface of a silica substrate. The two solutions were separately
2092 | Lab Chip, 2014, 14, 2090–2095
Fig. 3 A sketch of the in-fiber integrated CL optofluidic device. Inset: the structure of the connector.
contained with two injection pumps and connected with the connectors using polytetrafluoroethene (PTFE) capillaries (the outer diameter is 1.5 mm and the inner diameter is 0.5 mm). The PTFE capillaries were inserted into the connectors at the left end, the left inlet and outlet. They were immobilized using screws. The connectors were immobilized on the silica substrate using an adhesive. The solution can be injected into the optical fiber through the microholes without leakage. The inset of Fig. 3 presents the sketch of their connectivity to the optical fiber. At the beginning of CL reactions, light is emitted from the reagents and guided by the core of the optical fiber through evanescent field coupling39,40 of the core. The optical fiber provides enough length of the core for the collection of the light. Simultaneously, the light in the core can be detected at the remote end using a photomultiplier tube (PMT, R2949; Hamamatsu, Japan). The PMT is equipped with a high accuracy negative power source (biased at −1200 V for measurement) independently and connected with a photon counter. To avoid the interference of external light, the region within the red dashed box in Fig. 3 is placed in a dark box.
Results and discussion In this in-fiber integrated optofluidic CL analyzing system, two streams of fluids were pumped into the optical fiber. The first is a mixture of Vc, H2O2 and K3Fe(CN)6 and the other is luminol solution. K3Fe(CN)6 standard solution (1.0 × 10−2 M) was prepared by dissolving 0.330 g in a volumetric flask of 100 mL with distilled water. Vc standard solution was prepared by dissolving 3.649 g in a volumetric flask of 100 mL with distilled water. This solution was diluted to a specific concentration before use. H2O2 was used as the oxidant and prepared by diluting a commercial H2O2 solution with a concentration of 30% (w/w) to 0.4 mol L−1. Before injection, Vc, K3Fe(CN)6 and H2O2 solutions were mixed with the same volume. Luminol standard solution (1.0 × 10−2 M) was prepared by dissolving 0.272 g of luminol (Sigma-Aldrich) in a volumetric flask of 150 mL with NaOH solution (pH = 11). All of the
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 08 April 2014. Downloaded by University at Buffalo Libraries on 26/05/2014 15:47:38.
Lab on a Chip
standard solutions were stored at 4 °C in a refrigerator before use. When the above solutions were injected into the optical fiber, all of the streams were driven by syringe pumps at a fixed flow rate of 150 μl min−1 (the rate of each syringe pump is 75 μL min−1) which is an optimum value as discussed later. The typical spectrum of the CL reaction of luminol, K3Fe(CN)6, and H2O2 was obtained using a spectrometer (Fig. 4). From this curve, we can observe that this reaction has a typical emission of blue light and the maximum CL intensity is at around 430 nm. In this reaction, luminol is oxidized by H2O2 under the catalysis of K3Fe(CN)6. During the reaction, OH− removes the nitrogen protons, releasing a negative charge, which moves onto the carbonyl oxygen to form an enolate. Then, the oxidant (decomposed from H2O2 under the catalysis of K3Fe(CN)6) performs a cycloaddition reaction with the carbonyl carbons, forming an excited state (see the inset in Fig. 2) which emits blue light. Then, the photon is efficiently coupled into the inner core through the evanescent field and can be detected using the PMT. In this process, Vc influences the above reaction and the CL intensity depends on the concentration of Vc. To enhance the emission, the CL reaction was carried out in alkaline solution. The luminol solution was prepared in NaOH alkaline solutions with different pH values. Results indicate that CL reaction can produce the strongest emission in NaOH solution with a pH value of 11.0. Hence, this alkaline solution was chosen for the in-fiber integrated optofluidic analysis of Vc. When the two solutions were pumped hydrodynamically into the end and the opened inlet, the signal was remotely detected downstream of the point-of-confluence. The CL signal was observed to rise monotonically from zero to the maximum steady-state value within one second when the reagents were mixed. To present the change of the light intensity in the inner core when adjusting the concentration of Vc in the optofluidic reaction, the core at the end of the optical fiber was observed using a microscope. Fig. 5(a) shows that when the concentration of Vc increased from 0 to 2.81 mmol L−1, the intensity of the blue light at the end greatly decreased. The relationship between the CL intensity and the concentrations of Vc is presented in Fig. 5(b). The normalized CL
Fig. 4 The emission spectrum of the reaction of CL.
This journal is © The Royal Society of Chemistry 2014
Paper
Fig. 5 (a) Changes of the light intensity in the inner core during the optofluidic reaction. The concentration variation of Vc in the optical fiber: top left: 0 mmol L−1, top right: 0.65 mmol L−1, bottom left: 2.2 mmol L−1 and bottom right: 2.81 mmol L−1. (b) The relationship between the CL intensity and the concentration of Vc. Inset (top): the typical calibration plots for the Vc samples in the range of 0–3.0 mmol L−1. Inset (bottom): CL emission in the lateral view of the special optical fiber observed with a microscope.
intensity detected from the optical fiber with the air core is plotted against the concentration of Vc. From the response curve, we found that the intensity is obviously weakened when the concentration of Vc increases. The CL response of the device is nearly linear depending on the concentration of Vc in the range of about 0 to 3.0 mmol L−1 (inset (top) in Fig. 5(b)). In this range, the highest concentration of Vc is mainly determined by the lowest CL intensity which can be detected. Additionally, the inset (bottom) of Fig. 5(b) shows the CL emission in the lateral view of the special optical fiber observed with a microscope. Using the microscope, we can visually observe the emission change of the CL reaction with different concentrations of the samples. In this design, the intensity of the CL signal depends on the mixing dynamics within the optical fiber. At the optimal flow rate, reagents can be fully mixed by the time they reach the outlet, leading to a strong CL signal. To determine the optimal flow conditions for the experiments, we fixed the concentration of Vc at 0.2 mmol L−1 and investigated the influence of the total flow rate on the measured CL signal. The variation of the CL signal in response to changes of the flow rate is shown in Fig. 6. At slow flow rates such as 100 μl min−1,
Lab Chip, 2014, 14, 2090–2095 | 2093
View Article Online
Paper
Lab on a Chip
involving the process of CL or fluorescence quenching can be further integrated with this structure of the optical fiber.
Published on 08 April 2014. Downloaded by University at Buffalo Libraries on 26/05/2014 15:47:38.
Acknowledgements This work is supported by the National Natural Science Foundation of China (NSFC, 61007053, 61205027, 61377085, 61177081, 61227013), the 111 project (B13015) to the Harbin Engineering University, and the Fundamental Research Funds for the Central Universities.
Notes and references Fig. 6 Dynamic response of the in-fiber analyzing system.
the reaction is partially over within a short length of the optical fiber, leading again to a reduced signal. At higher flow rates such as 200 μl min−1, the reaction is inadequate by the time they reach the outlet and leads to a reduced CL signal. As a result, the strongest CL signal was obtained at an appropriate total flow rate of 150 μl min−1. In practice, the total flow rate was fixed at 150 μL min−1 to investigate the influence of Vc concentration on the intensity of the CL signal. Additionally, the dynamic response in Fig. 6 indicates that the steady-state CL emission can be obtained within 2 s. Conversely, when the sample injection is stopped, the emission quickly decreases and can also reach the minimum value within this time. These results also reveal that the sample solutions can form steady microflows around the core in the cave of the optical fiber.
Conclusions We describe a kind of optofluidic in-fiber integrated device based on a specially designed hollow optical fiber with an inner core. Especially, the path of the microfluid was built by etching microholes on the surface of this specially designed optical fiber without damaging the structure of the core which can be hardly realized in other kinds of MOFs. The results show that the CL reagents and samples can be mixed and form steady microflows in the optical fiber. Simultaneously, the emission from the reaction is efficiently coupled into the inner core through the long distance evanescent field, guided along the core through the melted point and then detected from the remote end of the optical fiber. Results indicate that the in-fiber integrated optofluidic device can achieve an efficient optical signal from the inner core while reducing the volume of the device. By analyzing the emission intensity of the fiber core, the linear sensing range of 0.1–3.0 mmol L−1 for Vc is obtained. The dynamic response reveals that the emission intensity can be determined within 2 s at a total flow rate of 150 μL min−1 (the rates of the two injection pumps are equal to 75 μL min−1). In addition, other in-fiber integrated optofluidic devices analyzing flowable substances such as some liquids and gases
2094 | Lab Chip, 2014, 14, 2090–2095
1 D. Psaltis, S. Quake and C. Yang, Nature, 2006, 442, 381–386. 2 X. D. Fan and S. Yun, Nat. Methods, 2014, 11, 141–147. 3 P. Müller, D. Kopp, A. Llobera and H. Zappe, Lab Chip, 2014, 14, 737–743. 4 I. Rodríguez-Ruiz, A. Llobera, J. Vila-Planas, D. Johnson, J. Gómez-Morales and J. García-Ruiz, Anal. Chem., 2013, 85, 9678–9685. 5 A. Mitra, B. Deutsch, F. Ignatovich, C. Dykes and L. Novotny, ACS Nano, 2010, 4, 1305–1312. 6 Z. Meng, X. Zhang and J. Qin, Nanoscale, 2013, 5, 4687–4690. 7 H. C. Fan, J. Wang, A. Potanina and S. R. Quake, Nat. Biotechnol., 2011, 29, 51–57. 8 K. H. Rasmussen, R. Marie, J. M. Lange, W. E. Svendsen, A. Kristensena and K. U. Mirb, Lab Chip, 2011, 11, 1431–1433. 9 J. Y. Liu and Y. X. Duan, Oral Oncol., 2012, 48, 569–577. 10 R. He, P. J. A. Sazio, A. C. Peacock, N. Healy, J. R. Sparks, M. Krishnamurthi, V. Gopalan and J. V. Badding, Nat. Photonics, 2012, 6, 174–179. 11 Y. Wang, D. Wang, C. Liao, T. Hu, J. Guo and H. Wei, Opt. Lett., 2013, 38, 269–271. 12 W. Qian, C. Zhao, Y. Wang, C. Chan, S. Liu and W. Jin, Opt. Lett., 2010, 34, 3296–3298. 13 X. H. Yang and L. Wang, Opt. Express, 2007, 15, 16478–16483. 14 Y. L. Ruan, H. Ebendorff-Heidepriem, S. Afshar and T. M. Monro, Opt. Express, 2010, 25, 26018–26026. 15 J. J. Kaufman, G. M. Tao, S. Shabahang, E. H. Banaei, D. S. Deng, X. Liang, S. G. Johnson, Y. Fink and A. F. Abouraddy, Nature, 2012, 487, 463–467. 16 Y. P. Wang, X. L. Tan, W. Jin, D. Q. Ying, Y. L. Hoo and S. J. Liu, Opt. Lett., 2010, 35, 88–90. 17 X. Yu, Y. C. Kwok, N. A. Khairudin and P. Shum, Sens. Actuators, B, 2009, 137, 462–466. 18 I. Al-Ogaidi, H. Gou, Z. Aguilar, S. Guo, A. Melconian, A. Al-kazaz, F. Mengand and N. Wu, Chem. Commun., 2014, 50, 1344–1346. 19 S. Casolari, B. Roda, M. Mirasoli, M. Zangheri, D. Patrono, P. Reschiglian and A. Roda, Analyst, 2013, 138, 211–219. 20 M. Kaczmarek and S. Lis, Anal. Methods, 2012, 4, 1964–1967. 21 B. Rezaei, A. Ensafi and F. Haghighatnia, Anal. Methods, 2012, 4, 1573–1578. 22 A. Natrajan, D. Sharpe and D. Wen, Org. Biomol. Chem., 2012, 10, 3432–3447.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 08 April 2014. Downloaded by University at Buffalo Libraries on 26/05/2014 15:47:38.
Lab on a Chip
23 K. Lok, S. Muttalib, P. Lee, Y. Kwok and N. Nguyen, Lab Chip, 2012, 12, 2353–2361. 24 D. J. Bowles, A. S. Olson, N. Ii, D. L. Heglund, R. L. Vargas and S. A. Williams, Appl. Spectrosc., 2012, 66, 175–17. 25 A. Stolyarov, A. Gumennik, W. McDaniel, O. Shapira, B. Schell, F. Sorin, K. Kuriki, G. Benoit, A. Rose, J. Joannopoulos and Y. Fink, Opt. Express, 2012, 20, 12407–12415. 26 N. Moghimi and K. T. Leung, Anal. Chem., 2013, 85, 5974–5980. 27 T. Huang, L. Yeh, H. Huang, Z. Nian, Y. Yeh, Y. Chou, J. Yeh and M. Tsai, Polym. Chem., 2014, 5, 630–637. 28 C. Sun, C. Chang, H. Lee, J. Zhou, J. Wang, T. Sham and W. Pong, ACS Nano, 2011, 5, 7788–7795. 29 L. Li, X. Cai, Y. Ding, S. Gu and Q. Zhang, Anal. Methods, 2013, 5, 6748–6754. 30 C. Weng, P. Hsu, S. Hsu, C. Chang, W. Hung, P. Wu and J. Yeh, J. Mater. Chem. B, 2013, 1, 4983–4991. 31 Q. Ma, Y. Li, Z. Lin, G. Tang and X. Su, Nanoscale, 2013, 5, 9726–9731.
This journal is © The Royal Society of Chemistry 2014
Paper
32 X. H. Yang, Y. Zheng, S. Luo, Y. Liu and L. B. Yuan, Sens. Actuators, B, 2013, 182, 571–575. 33 X. H. Yang, Y. Liu, Y. Zheng, S. Li, L. B. Yuan, T. Yuan and C. Tong, Opt. Commun., 2013, 304, 83–86. 34 A. Zhou, Y. Zhang, G. Li, J. Yang, Y. Wang, F. Tian and L. B. Yuan, Opt. Lett., 2011, 36, 3221–3223. 35 N. Mothe and P. Bin, Opt. Express, 2009, 17, 15778–15789. 36 J. Kim, J. Kim, Y. M. Jung, W. Ha, Y. Jeong, S. Lee, A. Tünnermann and K. Oh, Opt. Lett., 2009, 34, 2973–2975. 37 Y. J. Rao, Y. P. Wang, Z. L. Ran and T. Zhu, J. Lightwave Technol., 2003, 21, 1320–1327. 38 L. Lévesque and V. Jdanov, Opt. Laser Technol., 2010, 42, 1080–1083. 39 P. Lin, S. W. Kwok, H. Lin, V. Singh, L. Kimerling, G. Whitesides and A. Agarwal, Nano Lett., 2014, 14, 231–238. 40 Á. López-Lorente, M. Sieger, M. Valcárcel and B. Mizaikoff, Anal. Chem., 2014, 86, 783–789.
Lab Chip, 2014, 14, 2090–2095 | 2095
