We report a robust and sensitive optical nanofiber sensor with a femtoliter-scale detection volume. The sensor is fabricated by embedding a 800-nm-diameter nanofiber into a microfluidic chip with probing light propagated perpendicular to a 5-μm-wide detection channel. To verify the effectiveness of the sensor, we present measurements of fluorescence intensity and refractive index (RI) with detection limits of 1 × 10−7 M for fluorescein and 2.8 × 10−4 RIU, respectively. The femtoliter-scale optical nanofiber sensor shown here may provide a compact and versatile sensing platform for sensitive and fast detection of ultra-low-volume samples, as well as studying the dynamics of single molecule.
© 2015 Optical Society of America
With the advent of single-molecule detection and early disease diagnosis, there is a real need to reduce the detection volume with optical methods, because the concentrations of these samples are relatively high, but the amounts of these samples are quiet limit . To address this issue, total internal reflection with fluorescence correlation spectroscopy , confocal microscopy , zero-mode waveguides , and microstructured optical fibers  have enabled the observation and/or detection of molecules and ions in reduced volumes ranging from atto- to nanolitre-scale. All these approaches have strengths in different areas and are suitable for different applications. However, they require expensive instruments and/or complicated fabrication process, which prevent lots of researchers from studying and understanding biological processes on the molecular scale.
As an emerging sensing platform, in recent years, nanofiber-based optical sensing has been attracting increasing research interest due to its possibilities of realizing miniaturized fiber optic sensors with small footprint, high sensitivity, fast response, high flexibility and low optical power consumption [6–14 ]. Generally, light guided along an optical nanofiber leaves a large fraction of the guided fields outside the optical nanofiber as evanescent waves, while offers an opportunity to confine the majority of the evanescent fields within a wavelength-scale volume at the same time, which makes it highly sensitive to the highly localized index change of the surrounding medium , as well as offers a possibility to achieve a small detection volume by reducing the penetration depth of the evanescent field and the detection length of the nanofiber simultaneously. For example, by adjusting the diameter of the nanofiber and/or the wavelength of the probing light, the penetration depth of the evanescent field can be well confined within 100-200 nm, which is suitable for sensitive sensing with low background noise. Also, by embedding an optical nanofiber into a microfluidic chip with a detection channel perpendicular to the optical nanofiber, the detection length of the nanofiber that is defined by the width of the detection channel, can be precisely determined . To achieve a detection volume of ca. 1.0 femtoliter, a typical detection volume for single molecule analysis , it is necessary to reduce the channel width to less than 10 μm. However, the channel width of a typical microfluidic chip is in a range of tens to hundreds of micrometer . To obtain a sub-10-μm-wide channel, one has to use expensive instruments and time consuming protocols, such as combined electron beam and UV lithography , and femtosecond laser direct writing .
In our previous work, we presented a maskless method to fabricate nanochannels by using nanowires as templates , which can be easily adapted for fabricating sub-10-μm-wide channel by using micrometer-diameter microfibers as templates. More recently, by embedding optical nanofibers into a 125-μm-wide microchannel, we demonstrated evanescent wave absorption and fluorescence sensors that offer not only excellent stability and reproducibility, but also ultra-high sensitivity [21,22 ]. Herein, we report an optical nanofiber sensor by embedding a 800-nm-diameter nanofiber into a poly (dimethylsiloxane) (PDMS) microfluidic chip with a 5-μm-wide detection channel, resulting an effective detection volume down to ca.1 femtoliter. The performance of this sensing platform was investigated by measurements of fluorescence and refractive index (RI), achieving detection limits of 1 × 10−7 M for fluorescein and 2.8 × 10−4 RIU, respectively.
2. Experimental section
2.1 Fabrication of optical nanofibers, SU-8 microfibers, and SU-8 master
We fabricated optical nanofibers by stretching a standard single mode optical fiber while heating it with a hydrogen/oxygen flame . The typical transmission loss is about 1 dB. To obtain a microfluidic chip with a sub-10-μm-wide detection channel, we developed a novel method by using a SU-8 microfiber as a part of a SU-8 master to replicate PDMS microfluidic chips. The SU-8 microfibers were directly drawn from SU-8 photoresist on a glass slide as shown in Fig. 1(a) . A drop of SU-8 photoresist on the glass slide was heated to 95°C for 5 min, and a SU-8 microfiber was then drawn from the drop of the SU-8 photoresist by using a sharp tungsten probe. By adjusting the heating time or the viscosity of the SU-8 photoresist, the diameter of a SU-8 microfiber could be tuned from 1 to 10 μm. Figure 1(b) shows two SU-8 microfibers with diameters of 2 and 8 μm, respectively. To insure the surface gloss of as-fabricated SU-8 microfibers, the microfibers were collected by a home-made microfiber holder, and followed by exposure prior to the fabrication of SU-8 master. As the SU-8 microfibers were centimeters in length, we cut a section of microfiber with uniform diameter under a microscope for fabricating a SU-8 master.
To prepare a hybrid SU-8 master for replicating the PDMS microfluidic chips, a four-step fabrication procedure was developed. Firstly, a channel design without the detection channel on a mask was transferred onto a glass substrate. Secondly, a film of 10-μm-thick SU-8 photoresist was spin-coated on the substrate, followed by soft bake, exposure, post exposure bake and hard bake. Thirdly, a SU-8 microfiber was located onto the substrate by a 3D travel translation stage under an optical microscope. After post exposure bake and hard bake, the SU-8 microfiber attached tightly to the SU-8 film. Finally, a film of 130-μm-thick SU-8 photoresist was spin-coated on the substrate, followed by soft bake, expose from the bottom of the substrate, post exposure bake, developing, and hard bake.
2.2 Integration of an optical nanofiber with a microfluidic chip
In this work, we designed a microfluidic chip with sample inlet/outlet channels, fiber channels, and a narrow detection channel. In this case, the sample inlet/outlet channels (150 μm in width) can dramatically decrease the length of the detection channel, resulting in easier and faster sample loading/changing. With the help of the fiber channels (125 μm in width), the position of a nanofiber can be precisely controlled in the microfluidic chip, and the nanofiber will not bend at the position of detection channel. To integrate a nanofiber with a microfluidic chip, a PDMS slab with channels was mounted on a glass slide with the channels upside. A biconical tapered fiber was embedded into the fiber channels under an optical microscope. A PDMS membrane was then bonded to the PDMS slab by oxygen plasma treatment. Sample inlets and outlets, PDMS inlets were punched at the black circle points for sample loading/changing and PDMS injection, respectively (see Fig. 2(a) ). Importantly, to avoid sample leaking, uncured PDMS was carefully infused into the fiber channels from PDMS inlets, and the uncured PDMS was then diffused into the gap between the nanofiber and PDMS slab by capillarity. After curing at 65 °C for 30 min, only the part of the nanofiber across the detection channel can interact with sample solution, and the other part of the nanofiber was entirely embedded into the PDMS microfluidic chip. In this case, the microfluidic chip could provide small volume of sample to the nanofiber and protect the nanofiber from surface contamination and air disturbance. Also, we have tested the optical loss after integration. Typically, the optical losses were ca. 3 dB and 2dB for air and water filled the detection channel, respectively.
Figure 2(b) shows a typical optical micrograph of an as-fabricated nanofiber-microfluidic chip. A 500-μm-long, 8-μm-wide detection channel connects two 150-μm-wide channels for sample loading/changing, respectively. In order to further reduce the width of the detection channel, a 5-μm-diameter SU-8 microfiber was used to replicate the detection channel. Figure 2(c) shows the side view of a 5-μm-wide channel, and the blue dot line indicates the position of the nanofiber. Because the SU-8 microfiber partially merged with the SU-8 substrate after post exposure bake and hard bake, the cross section of the detection channel was not a complete circle. The estimated channel width at the top of the channel is ca. 3.8 μm. Furthermore, the PDMS might cover a small part of the nanofiber when we introduced PDMS into the fiber channels to avoid sample leaking. When 0.01 mM fluorescein solution was introduced into the 5-μm-wide detection channel, and a 473-nm-wavelength laser was launched to the nanofiber, a bright fluorescence spot excited by the evanescent field outside the nanofiber can be seen in Fig. 2(d), indicating an effective detection length of 2.5 μm and no sample leaking.
3. Results and discussion
3.1 Modling the power distribution outside the nanofiber
Figure 3(a) shows power distribution (Z-direction Poynting vectors) of HE11 mode of a water cladding 800-nm-diameter silica nanofiber operated at 473-nm-wavelength in 3D view. It is clear that, the 800-nm-diameter nanofiber confines major energy inside the fiber, and leaves ~11% of light guided outside as evanescent waves (see Fig. 3(b)). When we defined the penetration depth as the length where the evanescent field intensity decays to 10% of the highest intensity outside the nanofiber , the penetration depth is calculated to be about 150 nm (see Inset of Fig. 3(b)), leading to an effective detection volume of ca. 1.0 femtoliter for a detection length of 2.5 μm.
3.2 Measurement of fluorescence intensity
Fluorescence measurements provide sensitive detection in biochemical analysis, immunoassay, and single molecule detection. Recent research works have demonstrated the feasibility and potential of using nanofibers as fluorescence sensors for chemical and biological applications [22,24,25 ]. Fluorescein is a fluorophore commonly used in microscopy, it has an absorption maximum at 494 nm and emission maximum at 512 nm in water. In this work, we investigated the sensitivity and linearity of the sensor by measuring fluorescence intensity of fluorescein solutions with concentrations ranging from 1 × 10−7 to 1 × 10−6 M, which were prepared before use. 1 μL of sample solution was injected into one of the sample inlets, and the sample solution was then introduced into the sample inlet channel by exerting negative pressure at the other sample inlet. In order to introduce the sample into the 5-μm-wide channel from the sample inlet channel, one of the sample outlets was covered by a piece of PDMS membrane and negative pressure was generated by a syringe at the other sample outlet. After each measurement, the microchannels were flushed with ultrapure water.
To obtain fluorescence intensity, we took the optical micrographs of the fluorescence by a charge coupled device (CCD) camera mounted on a microscope. The colorful photograph was converted to gray scale photograph, a section of 50 × 25 pixel gray scale photograph was then selected for gray value calculation. The gray value for each pixel is ranging from 0 to 255, the gray values for all the pixels in the selected section were added together as the fluorescence intensity by a self-compiled MATLAB program. When the concentration of fluorescein (CFluorescein) increased from 1 × 10−7 to 1 × 10−6 M, the intensity of fluorescence increases obviously (Insets (a-e) of Fig. 4 ). A linear concentration-dependent response of Gray value (fluorescence intensity) = 44537CFluorescein (10−7 M) – 7426 was obtained as shown in Fig. 4. Theoretically, the concentration resolution can be 2 × 10−11 M based on the variation of gray value of 1. Note that a high resloution CCD is favorable for high concentration resolution because more pixels will contribute to gray value calculation. Since the fluorescence intensity is proportional to excitation intensity, lower or higher concentration sample solutions could be measured by adjusting laser power, exposure time, and diameter of the nanofiber.
3.3 Measurement of refractive index
RI is extremely useful for label free sensing and detecting compounds without absorbance in the UV-vis range or fluorescence . In this work, we used an unpolarized broadband light from a tungsten halide lamp as a probing light owing to its excellent stability of beam intensity. The probing light was coupled into the nanofiber, and the transmitted light was coupled into a spectrometer (Maya2000 Pro, Ocean optics, Dunedin, FL, USA) as signal for real time RI sensing. A series of ethylene glycol solutions with RI ranging from 1.335 to 1.405 were prepared before use. Figure 5(a) schematically shows the cross section of the detection channel and the nanofiber embedded in the PDMS microfluidic chip. When there is a RI contrast between the sample in the detection channel and the PDMS channel wall, the probing light will be strongly scattered at the channel boundaries, the larger RI contrast, the stronger scattering. When the RI of a sample (ns) matches the RI of PDMS (ca. 1.40), the probing light confines mostly in the nanofiber, resulting in the maximal transmission. Since ethylene glycol solutions have no absorbance in the UV/vis range, the contribution from the absorbance of the liquid is negligible compared with the scattering.
As shown in Fig. 5(b), with the increase of RI, the transmission increased obviously. The transmission is at maximum when the RI of the sample (ns = 1.405) matches the RI of PDMS, and we defined it as 100% transmission, thus, the sensor’s working range depends on the RI of the detection channel. Note that the longer wavelength is more sensitive to RI variations in the detection channel. Similar to the effect of the reducing diameter at a certain wavelength , evanescent fields moves out of the core of the fiber drastically when the wavelength increase to a critical value. For a 800-nm-diameter nanofiber operated at 900-nm-wavelength, a majority of the light is moved out of the nanofiber and propagated as evanescent waves. When RI changes from 1.345 to 1.395, the transmission of the sensor operated at 900-nm-wavelengh increased 70%. Because the stability of the light source is measured with an outstanding reproducibility of 0.13% RSD, the RI sensitivity in the range from 1.345 to 1.395 can be estimated as 2.8 × 10−4 RIU based on 3 times the standard deviation of the light source. This sensitivity exceeds that of three-dimensional photonic crystal RI sensor  and comparable with that obtained using the evanescent field-based optical fiber RI sensor, which typically consume 100-fold more sample . When thinner nanofiber operates at a longer wavelength, it should be possible to reach a higher sensitivity. The reversible response of the sensor was tested by cycling between water and an ethylene glycol/water mixture (n = 1.345) in the detection channel, and the detection wavelength was set at 900 nm. The data in the inset of Fig. 5(b) confirmed an excellent reversibility. We estimated a response time to be about 600 ms for the sensor with sample throughput of 90 h−1. This fast response and high throughput can be attributed to the microchannel networks that enable rapid sample changing and small sample consumption.
In summary, we have demonstrated a femtoliter-scale nanofiber sensing platform by taking the advantages of the tightly confined large fractional evanescent fields of the waveguiding nanofiber, as well as the short detection length defined by the width of a narrow microfluidic channel. When a 800-nm-diameter nanofiber integrated with a microfluidic chip with a 5-μm-wide detection channel, the effective detection volume can be down to ca. 1.0 femtoliter. We investigaed the performance of the nanofiber-microfluidic sensor by measuring fluorescence intensity and RI, achieving detection limits of 1 × 10−7 M for fluorescein and 2.8 × 10−4 RIU, respectively. Although the detection limits shown here are worse than our proir work owing to less interaction length and evanescent wave strength, the femtoliter-scale detection volume is attractive for high concentration biological samples with small volumes. When an electron-multiplying charge-coupled device (EMCCD) or a fast response, high resolution spectrometer is used to record the signal, the femtoliter-scale optical nanofiber sensor shown here may provide a compact and versatile sensing platform for sensitive and fast detection of ultra-low-volume samples, as well as studying the dynamics of single nanoparticles such as single molecule.
This work was supported in parts by the National Basic Research Program of China (Nos. 2013CB328703 and 2014CB921303), the National Natural Science Foundation of China (NSFC) (Nos. 61275217, 21205109 and 21407039), and National Science and technology support program (No. 2012BAK08B05).
References and links
2. T. E. Starr and N. L. Thompson, “Total internal reflection with fluorescence correlation spectroscopy: combined surface reaction and solution diffusion,” Biophys. J. 80(3), 1575–1584 (2001). [CrossRef] [PubMed]
4. M. J. Levene, J. Korlach, S. W. Turner, M. Foquet, H. G. Craighead, and W. W. Webb, “Zero-mode waveguides for single-molecule analysis at high concentrations,” Science 299(5607), 682–686 (2003). [CrossRef] [PubMed]
5. S. Heng, M. C. Nguyen, R. Kostecki, T. M. Monro, and A. D. Abell, “Nanoliter-scale, regenerable ion sensor: sensing with a surface functionalized microstructured optical fibre,” RSC Advances 3(22), 8308–8317 (2013). [CrossRef]
7. P. Polynkin, A. Polynkin, N. Peyghambarian, and M. Mansuripur, “Evanescent field-based optical fiber sensing device for measuring the refractive index of liquids in microfluidic channels,” Opt. Lett. 30(11), 1273–1275 (2005). [CrossRef] [PubMed]
11. P. Wang, G. Brambilla, M. Ding, Y. Semenova, Q. Wu, and G. Farrell, “High-sensitivity, evanescent field refractometric sensor based on a tapered, multimode fiber interference,” Opt. Lett. 36(12), 2233–2235 (2011). [CrossRef] [PubMed]
14. X. C. Yu, B. B. Li, P. Wang, L. Tong, X. F. Jiang, Y. Li, Q. Gong, and Y. F. Xiao, “Single nanoparticle detection and sizing using a nanofiber pair in an aqueous environment,” Adv. Mater. 26(44), 7462–7467 (2014). [CrossRef] [PubMed]
16. D. J. Sirbuly, A. Tao, M. Law, R. Fan, and P. Yang, “Multifunctional nanowire evanescent wave optical sensors,” Adv. Mater. 19(1), 61–66 (2007). [CrossRef]
19. B. B. Xu, Y. L. Zhang, H. Xia, W. F. Dong, H. Ding, and H. B. Sun, “Fabrication and multifunction integration of microfluidic chips by femtosecond laser direct writing,” Lab Chip 13(9), 1677–1690 (2013). [CrossRef] [PubMed]
20. L. Zhang, F. Gu, L. Tong, and X. Yin, “Simple and cost-effective fabrication of two-dimensional plastic nanochannels from silica nanowire templates,” Microfluid. Nanofluidics 5(6), 727–732 (2008). [CrossRef]
23. W. Li, B. Chen, C. Meng, W. Fang, Y. Xiao, X. Li, Z. Hu, Y. Xu, L. Tong, H. Wang, W. Liu, J. Bao, and Y. R. Shen, “Ultrafast all-optical graphene modulator,” Nano Lett. 14(2), 955–959 (2014). [CrossRef] [PubMed]
24. A. Stiebeiner, O. Rehband, R. Garcia-Fernandez, and A. Rauschenbeutel, “Ultra-sensitive fluorescence spectroscopy of isolated surface-adsorbed molecules using an optical nanofiber,” Opt. Express 17(24), 21704–21711 (2009). [CrossRef] [PubMed]
25. S. C. Warren-Smith, S. Afshar, and T. M. Monro, “Fluorescence-based sensing with optical nanowires: a generalized model and experimental validation,” Opt. Express 18(9), 9474–9485 (2010). [CrossRef] [PubMed]
28. J. Wu, D. Day, and M. Gu, “A microfluidic refractive index sensor based on an integrated three-dimensional photonic crystal,” Appl. Phys. Lett. 92(7), 071108 (2008). [CrossRef]