Open Access
Add to BibSonomyAbstract
An optical fiber containing longitudinal holes adjacent to the core has been used to detect and collect fluorescent particles from a solution. Excitation light was launched through the fiber and fluorescence signal was guided back to a detector system. As a proof of principle, green and red fluorescent polystyrene beads were detected and selectively collected from a water solution containing a mixture of red and green fluorescent beads.
© 2014 Optical Society of America
1. Introduction
Bacterial quantification by counting is required in various disciplines including water treatment, public health and pharmaceutical industries [1]. Traditional bacterial quantification methods include culture-based techniques [2,3] and flow cytometry [4,5]. Culture-based techniques are considered to be time consuming and do not provide real-time information, while flow cytometry has complex instrumentation and is costly. Quantification and isolation of biological cells (other than bacteria) are also required in medicine, e.g., for diagnosis of breast cancer [6] and for prenatal diagnosis [7]. Traditional cell sorting methods based on fluorescence [8] and immunomagnetic separation [9] are rapid but often require milliliter samples. Lab-on-a-chip and other microdevices provide opportunity for cell and bacterial quantification using picoliter to nanoliter sample volumes [10], for high throughput separation [11] and isolation within a short time frame [12]. An unresolved challenge is that the measurements are often performed discontinuously, e.g., by extraction of sample by needle biopsy. Optical fibers, on the other hand, can be inserted into different non-transparent environments and areas of difficult access. They open the possibility for minimally-invasive real-time monitoring in-vivo [13]. It is of interest to add new functionality to optical fibers, for example to allow for their use in identifying and retrieving small particles. This may find future applications in life-science.
In this work we present the proof-of-principle of a method to detect and collect micrometer-size particles with a fiber optic probe and laser-induced fluorescence (LIF). LIF detection with optical fibers has previously been used, for instance, to measure diffusion in environmental microbiology [14]. The fiber-based detection system described below allows for real-time monitoring and for collection of the detected particles. The light launched through the fiber into the sample volume excites particles in the proximity of the fiber-tip. The fluorescence returns through the fiber to the detector, triggering a pump that sucks the particles through longitudinal holes in the fiber. This technique finds applications in the selection of rare particles in fluids and for in-vivo collection of cells. The LIF detection method used here could potentially be exchanged or combined with Raman or other scattering detection methods.
2. Measurement principle
The technique introduced below relies on exciting particles in a solution with laser light and detecting their fluorescence. An optical microstructured fiber is used for excitation and detection. When the fluorescence signal exceeds a predetermined threshold, the particle is considered to be of interest and to be sufficiently near the fiber-tip to be captured. A suction mechanism is then activated, which brings the particle into the hole in the microstructured fiber. Several aspects deserve a closer look to make the technique useful. It is assumed in the following that the particles reside in a transparent fluid with no optical loss from absorption or scattering.
2.1 Excitation of on-axis particles
Consider a fiber with known numerical aperture (NA) n0sinα in the fluid (with fluid refractive index, n0), mode field radius R0 and delivering an intensity I0 at the center of the core at the fiber-tip. Since the light spreads in the fluid as a cone, particles along the fiber axis and at a distance, L, from the fiber-tip are excited by an intensity, I1, related to I0 through I1/I0 = {1 / (1 + L tanα / R0) }2. In the application below, tanα = 0.10 in water (0.13 in air) and R0 ~4 µm.
2.2 Fluorescence detected from on-axis particles
Excited particles fluoresce and the emission is uniform in a sphere of solid angle 4π sr. Consider again the case when the particle is aligned along the fiber axis and its distance from the fiber-tip is L. It can be shown that the fraction of the light collected by a core of mode-field radius R0 is then η = { πR02 + π[(R02 + L2)1/2 – L]2 } / 4π(L2 + R02). When L >> R0, this expression reduces to R02/4L2, i.e., the fraction of the light collected falls with the square of the distance from the particle to the fiber-tip. The combined excitation and signal detection for particles relatively far away from the fiber-tip falls with L4, and is rigorously described by the radar equation [15].
In the important case when the distance between the particle and fiber is comparable with the core radius, one needs to consider that the collection becomes constant for angles exceeding the NA of the fiber, even if the particle is very close, almost touching the core. This angle is 5.7° in water for the fiber used here and with R0 = 4 µm, the collection efficiency becomes constant for L ≤ 40 µm. The amount of collected light is η = 2.5 × 10−3. For shorter distances between particle and fiber-tip (L < 40 µm), the fraction of the collected light stays at a level η = 2.5 × 10−3.
2.3 Off-axis particles
In sections 2.1 and 2.2, it is assumed that the particle lies along the fiber axis. To a good approximation, the far-field intensity distribution of the fundamental mode is Gaussian. When the particle is displaced sideways from the axis by r, the optical coupling reduces by ~exp(-r2) both for the excitation of the particle and for the collection of its fluorescence. This leads to a strong reduction of the optical signal detected from the off-axis particle at a plane distance L from the fiber-tip, when compared to the intensity detected from an on-axis particle on the same plane, implying that the detection of particles very near the axis is strongly favored, compared to off-axis particles.
The graph in Fig. 1 illustrates on a log-log scale the calculated fraction of the signal detected through the fiber as a function of the distance between the particle and the fiber-tip, for on-axis particles, in the case of unit efficiency of emission and detection. The blue curve describes the excitation, the red one illustrates the detection and the black one combines excitation and detection. One sees that for distances greater than ~100 µm the L−4 dependence is observed. The signal detected is flattened from the fact that for distances < 40 µm the NA is exceeded and the collection efficiency saturates. One also notes that when the particle distance increases from 40 µm to 100 µm, the optical signal detected reduces strongly (by 20 times). The signal detected from particles 1 mm away from the fiber-tip are 4 orders of magnitude weaker than that from particles 100 µm away from the fiber-tip. For the off-axis particles, the detected signal would be increasingly weaker as the particle is moved further from the axis.
Fig. 1 Calculation of the signal detected for the arrangement implemented experimentally below. The blue curve shows the excitation signal reaching on-axis particles, the red curve the fraction of the signal detected for unitary efficiency of emission and detection, and the black curve the signal detected taking into account the two effects above.
2.4 Collection of particles
In the ideal case, the present optically-triggered particle collection technique allows retrieving the particles of interest with the suction of a negligible volume of fluid. In reality, however, a minimum volume of liquid needs to be collected to guarantee that the particle that triggers the signal is indeed sucked into the hole of the microstructured fiber. An estimate of this minimum volume is calculated in the following way. As seen above, the optical fluorescence signal measured depends very strongly on the distance between the particle and the fiber-tip. The ability to trigger the system to initiate collection drops rapidly for particles further than ~40 µm from the tip, as well as for particles well off-axis. Referring to Fig. 2, let D be the distance between the center of the core and of the hole used for suction, L be the distance between the fiber-tip and an on-axis particle capable of triggering the system and R the distance between the particle and the center of the hole. Assume the liquid not to be viscous. For small D and L it suffices to retrieve a volume amounting to ~2/3 πR3 = 2/3 π(L2 + D2)3/2 to ensure that the particle is collected.
Fig. 2 In order to guarantee the collection of an on-axis particle that triggers the suction system distance L from the fiber-tip and R from the hole entrance, a minimum volume of fluid needs to be retrieved ~2/3 πR3. This volume should be minimized for improved performance.
It is clear that by design, the distance D between core and hole should be as small as possible to minimize the volume of liquid accompanying the particle of interest. Likewise, the collection system performs better if the trigger level is set sufficiently high to make sure that L is very small for a trigger event to take place (at the expense of waiting longer between triggered events).
In the implementation below, light guidance and fluid collection take place in a single fiber. The distance, D, between the fiber core and the ~25-30 µm diameter hole is ~34 µm. Had the microstructured fiber been replaced by a standard fiber and a neighboring 125-µm outer diameter capillary, D would have been 125 µm, i.e., 3.7 times larger than achieved here with the use of a microstructured fiber. From the expression for the volume above, a 3.7 times increase in D corresponds to an increase in the volume accompanying the particle of 50 times. This justifies the use of a microstructured fiber rather than the combination of a fiber for excitation and a capillary for collection.
When dilution is allowed (e.g., in vitro), one could dilute the sample to a point when the collection of wanted species is not associated with the collection of unwanted species (i.e., specificity near 100% at the expense of longer collection time). However, if dilution is not allowed (e.g., the system is to be used in-vivo and fluid should not be injected) the only way to increase the specificity of the sample collected is to minimize the volume of fluid that accompanies a triggered event. In this case, it is crucial to reduce the minimum volume collected. Keeping the separation core-hole (D) small becomes imperative.
3. Experiments
The experimental system lay-out can be seen in Fig. 3. A compact, low noise, sum-frequency mixed diode-pumped solid-state laser emitting at 491 nm wavelength was used for excitation [16]. A ~1 m long piece of standard telecom fiber (SMF 28) was fused to a ~24 cm long microstructured fiber with longitudinal holes adjacent to the core, which was used to deliver light to the sample volume and to collect detected particles. The excitation beam was launched into the fiber core through a dichroic beam splitter and a 10x focusing lens, and propagated through the fiber to its tip. The output power at the fiber tip was typically ~3 mW. The fluorescence signal from excited particles was collected and guided backwards through the fiber, transmitted through the dichroic beam-splitter and filters and detected by a 50-Ω impedance photomultiplier tube (PMT). The electrical signal was digitized in a 16 bit A/D converter and taken to the computer for trigger-decision. An oscilloscope was used in parallel to visualize the signal. Backward-scattered laser radiation and other stray light were filtered out using sharp interference filters and color filters placed before the PMT, to ensure that only the fluorescence from the particles was detected. At an early stage of the project it was noticed that fluorescence from the fiber acrylate coating contributed to an increase in background noise. Replacing the SMF28 with a fiber provided with a ~20-nm thick carbon-coating on the cladding surface, decreased the background noise and, thus, increased the signal-to-noise ratio 2-3 orders of magnitude [17]. A solenoid valve was connected to the vacuum pump used with the fiber arrangement, as illustrated in Fig. 3. The valve was controlled with a LABVIEW program to automatically open/close when the fluorescence signal exceeded the trigger level, i.e. indication of a particle to be collected.
Microstructured fibers with either one or two holes were used for the experiments, as seen in Figs. 4(a) and 4(b). Both fiber types had outer diameter 125 µm and core diameter 8 µm. The 1-hole fiber had a slightly wedged hole with diameter ~30 µm, while the 2-hole fiber had ~26 μm diameter holes. The distance from the edge of the fiber hole to the core was 18 µm. The 2-hole fiber is symmetric and allows for low-loss splicing, but for particle collection, only one hole was open to permit flow, referred to in the following as the “active” hole.
Fig. 4 SEM-images of the 125-μm diameter microstructured fibers used in this work. (a) Illustration of the 1-hole fiber with a ~30 µm diameter hole and (b) the 2-hole fiber with ~26 µm diameter holes. Both fibers had 8-μm cores.
As mentioned above, the microstructured fiber was spliced to a carbon-coated fiber without openings. In order to allow for liquid flow through the microstructured fiber, the active hole was opened from the side by polishing, approximately 20 cm away from the fiber tip. The polishing was done using a 2400 mesh paper attached to a rotating drum. A 20-µm thin metal wire was inserted into the hole before polishing to avoid debris from clogging the hole. The metal wire was removed after polishing. The splice loss between the 2-hole fiber and the carbon-coated fiber (CCF) was ~0.1 dB and the corresponding loss for the 1-hole fiber was ~1dB. The spliced fiber was then inserted into an 8 cm long 0.5 mm inner-diameter needle from the side, through an opening cut open in the metal by polishing, 3 cm from the needle’s tip. The metallic needle served three purposes: as a pressure chamber to allow pumping the microstructured fiber through its opened active hole; as an arrangement for fiber optical coupling through the side-passage cut on the metal wall; and as a packaging means to protect the fibers, the splices and the fragile side-polished fiber, positioned inside the needle. Both fiber and needle were accommodated inside a conventional 6 cm splice protector, as illustrated in Fig. 5. The needle could then be easily coupled to a vacuum pump, to allow for sample collection in the fiber using under-pressure. In this way, liquid flow was allowed in the microstructured fiber without disturbing the optical light guidance and minimizing risk of leakage or contamination. The 2-hole fiber tip was finally cleaved to obtain a clean and smooth end-face.
Fig. 5 Image of carbon-coated fiber (on the left) spliced to a microstructured fiber (on the right) which had been side-polished to open the active hole. This opening in the fiber is placed inside a needle, that works as a pressure vessel, and which is protected by a conventional splice protector. The needle itself is side-polished to allow for lateral entrance of the carbon-coated fiber. The opening in the metal is sealed by the splice protector.
4. Results and discussion
Proof-of-principle experiments were performed in which the 2-hole fiber was used. Fluorescent polystyrene beads (Thermo Scientific Fluoro-Max Green Aqueous Fluorescent Particles) of diameter 5 µm and emission at 508 nm and 6 μm Polysciences beads with emission at 630 nm were detected and collected from liquid solutions. A 50-µl volume sample was put in a small container and the fiber-tip was then inserted into the container. The sample container was vibrated to prevent the beads from sedimenting and improve uniformity. When the beads were within a distance of ~40 µm away from the fiber-tip, the S/N ratio was up to 50/1. The detection trigger level was set to an intensity level adjusted to ensure that only beads within a reasonably small distance (e.g., ~40 µm) triggered the opening of the valve. Since beads of different fluorescence efficiency were used, this level was adjusted empirically.
Figures 6 and 7 illustrate the collection and sorting of beads with accompanying film sequences. The videos show the fiber inserted in a liquid with fluorescing beads. The shaker used previously to keep particles in movement was switched off to allow recording with a stationary fiber-tip under the microscope. The first video (Media 1) shows green bead collection in a viscous liquid (Mounting media, InSpeck, Lifetechnologies). A graph attached to the sequence also illustrates the signal generated by the photomultiplier in real time and the constant signal level chosen for trigger. When a particle appears off-axis or is far from the fiber-tip, the fluorescence signal detected is insufficient to exceed the trigger level. When the particle’s movement in the liquid brings it nearer to on-axis excitation (or as shown here, when the fiber is moved to a sufficiently favorable excitation level) the PMT signal increases, and exceeds the trigger level, starting the pumping mechanism. Although a sound track is not included, the activation of the suction is heard with a click, and starts with the trigger event. In the first sequence (Media 1), the process purposely happens very slowly. This allows one to follow the particle retrieved all the way into the fiber hole. One notices even that the particle becomes darker as suction brings it into the hole (no longer facing the exciting laser beam). The corresponding pictures illustrate a green fluorescence particle facing the fiber core at the activation instant (Fig. 6(a)) and the same particle already collected and traveling in the fiber hole (Fig. 6(b)).
Fig. 6 (See Media 1.) Collection of green fluorescent beads. (a) The fluorescence signal produced by the particle and detected activates the trigger and starts the suction mechanism. (b) The bead is retrieved into the hole with the surrounding liquid volume.
Fig. 7 (See Media 2.) (a) Time sequence of the detected signal in the PMT showing that three trigger events take place. Three green fluorescent beads are collected. (b) Illustration of fluorescent green bead 42 µm from the fiber-tip, just before collection. (c) Time sequence of the detected signal during collection of red beads. One particle, shown in (d) is collected.
The second video (Media 2) shows the collection of particles in water. Here, suction happens so fast that it is difficult to see the particle’s movement into the hole. The sequence shows that it is possible to choose between collection of green and red fluorescent beads. The type of particle collected depends on the choice of filter in front of the photomultiplier. Figure 7(a) shows the time sequence of the signal detected during the collection of green beads, where the trigger level 5 × 10−2 (a.u.) is exceeded three times (three beads are collected). Figure 7(b) illustrates the position of a green bead (~42 µm from the fiber-tip) immediately before collection is initiated. Figure 7(c) shows the time sequence of the PMT signal when red beads are chosen. Since fluorescence is weaker, the trigger level is lowered to 5 × 10−3 a.u. A single collection event is initiated during this sequence. Figure 7(d) shows the position of the red fluorescence bead (39 µm away from the fiber tip) just before collection is initiated.
The ideal collection retrieves every bead that generates a trigger signal and a minimum of fluid with it. Preferably, one and only one bead should be collected at a time. This sets an upper limit on the allowed concentration of beads in the solution. This maximum concentration is estimated below, assuming that the beads are uniformly distributed. As seen in section 2.4, the minimum volume collected to guarantee that the bead is collected at every triggered event is equal to 2/3 (πD2). For D ~34 µm, this corresponds to a collection volume of ~300 pl. Under typical experimental conditions used here, the valve was opened for 40 ms and a liquid volume of ~300 pl was sucked into the active hole of the fiber. In this case, the sample concentration should not exceed ~3·106 beads/ml to guarantee single bead collection per trigger event.
5. Conclusions and outlook
We have demonstrated a fiber-based LIF detection system which can be used to identify single micrometer-size particles from both homogeneous (sampling) and heterogeneous (separation) sample solutions, and collect them with the hole integrated in the fiber. The method presented allows for real-time monitoring of the collection of particles, which is an advantage when it comes to in-vivo measurements.
In the case where dilution of the sample is allowed, through particle selection one can enrich the sample concentration up to the level one particle per collected volume in the fiber and obtain a specificity near to 100%.
With the 2-hole fiber it should be possible to sort two types of particles with different fluorescence signature, one in each hole. It should also be possible to utilize the holes for injection of small amounts of chemical or biological substances on demand, thereby inducing chemical and biological reactions combined with local photochemistry. Further work could also include development of microstructured fibers with several integrated holes in a more complex system, where each individual hole can have a specific function i.e., collection or delivery of different substances.
Acknowledgments
This work was supported by the Swedish Research Council (VR) through its Linnæus Center of Excellence ADOPT. We are also very grateful to Prof. Gunnar Björk, Prof. Gustav Amberg, and Dr. Aman Russom at the Royal Institute of Technology for their valuable input and help. The special fibers used in this work were made at Acreo Swedish ICT Fiberlab.
References and links
1. O. Lazcka, F. J. Del Campo, and F. X. Muñoz, “Pathogen detection: a perspective of traditional methods and biosensors,” Biosens. Bioelectron. 22(7), 1205–1217 (2007). [CrossRef] [PubMed]
2. P. B. Crone, “The counting of surface colonies of bacteria,” J. Hyg. (Lond.) 46(4), 426–430 (1948). [CrossRef] [PubMed]
3. M. Bölter, J. Bloem, K. Meiners, and R. Möller, “Enumeration and biovolume determination of microbial cells - a methodological review and recommendations for applications in ecological research,” Biol. Fertil. Soils 36(4), 249–259 (2002). [CrossRef]
4. D. A. Veal, D. Deere, B. Ferrari, J. Piper, and P. V. Attfield, “Fluorescence staining and flow cytometry for monitoring microbial cells,” J. Immunol. Methods 243(1-2), 191–210 (2000). [CrossRef] [PubMed]
5. L. Yang, Y. Zhou, S. Zhu, T. Huang, L. Wu, and X. Yan, “Detection and quantification of bacterial autofluorescence at the single-cell level by a laboratory-built high-sensitivity flow cytometer,” Anal. Chem. 84(3), 1526–1532 (2012). [CrossRef] [PubMed]
6. M. Cristofanilli, G. T. Budd, M. J. Ellis, A. Stopeck, J. Matera, M. C. Miller, J. M. Reuben, G. V. Doyle, W. J. Allard, L. W. Terstappen, and D. F. Hayes, “Circulating tumor cells, disease progression, and survival in metastatic breast cancer,” N. Engl. J. Med. 351(8), 781–791 (2004). [CrossRef] [PubMed]
7. K. Krabchi, F. Gros-Louis, J. Yan, M. Bronsard, J. Massé, J. C. Forest, and R. Drouin, “Quantification of all fetal nucleated cells in maternal blood between the 18th and 22nd weeks of pregnancy using molecular cytogenetic techniques,” Clin. Genet. 60(2), 145–150 (2001). [CrossRef] [PubMed]
8. H. M. Davey and D. B. Kell, “Flow cytometry and cell sorting of heterogeneous microbial populations: the importance of single-cell analyses,” Microbiol. Rev. 60(4), 641–696 (1996). [PubMed]
9. J. J. Chalmers, M. Zborowski, L. Sun, and L. Moore, “Flow through, immunomagnetic cell separation,” Biotechnol. Prog. 14(1), 141–148 (1998). [CrossRef] [PubMed]
10. N. Bao, B. Jagadeesan, A. K. Bhunia, Y. Yao, and C. Lu, “Quantification of bacterial cells based on autofluorescence on a microfluidic platform,” J. Chromatogr. A 1181(1–2), 153–158 (2008). [CrossRef] [PubMed]
11. Z. Wu, B. Willing, J. Bjerketorp, J. K. Jansson, and K. Hjort, “Soft inertial microfluidics for high throughput separation of bacteria from human blood cells,” Lab Chip 9(9), 1193–1199 (2009). [CrossRef] [PubMed]
12. S. J. Tan, L. Yobas, G. Y. H. Lee, C. N. Ong, and C. T. Lim, “Microdevice for the isolation and enumeration of cancer cells from blood,” Biomed. Microdevices 11(4), 883–892 (2009). [CrossRef] [PubMed]
13. S. Sato, T. Ando, and M. Obara, “Optical fiber-based photomechanical gene transfer system for in vivo application,” Opt. Lett. 36(23), 4545–4547 (2011). [CrossRef] [PubMed]
14. M. Kühl, Methods in Enzymology (Elsevier, 2005), Chap. 10.
15. E. F. Knott, J. F. Schaeffer, and M. T. Tuley, Radar Cross Section (SciTech Publishing, 2004), Chap. 2.
16. M. Stjernström, F. Laurell, and H. Brismar, “Diode-pumped solid state laser light sources for confocal laser scanning fluorescence microscopy,” J. Laser Appl. 20(3), 160 (2008). [CrossRef]
17. A. Sudirman, L. Norin, and W. Margulis, “Increased sensitivity in fiber-based spectroscopy using carbon-coated fiber,” Opt. Express 20(27), 28049–28055 (2012). [CrossRef] [PubMed]
