Emerald Group Publishing Limited
Copyright © 2011, Emerald Group Publishing Limited
Article Type: Nanosensor update From: Sensor Review, Volume 31, Issue 4
Sensing with optical nanofibres
As illustrated elsewhere in this issue, fibre optic sensors have achieved considerable commercial success but many believe that one of the most effective means of achieving improved performance and wider applications lies in the miniaturisation of the sensing structure. Accordingly, this update considers the sensing implications of reducing an optical fibre to nanometre dimensions. This can take either of two forms: reducing the diameter of a section of fibre or reducing that of the fibre tip. Sensor developments based on of both of these concepts are discussed.
Characteristics of optical nanofibres
Optical nanofibres (ONFs), also termed photonic nanowires or sub-wavelength fibres, are optical fibres with diameters ranging from tens of nanometres to a few hundred nanometres (Figure 1). The diameter is therefore often well below the optical wavelength and such fibres exhibit interesting mechanical and optical properties. Owing to the large difference in refractive index between the fibre and air, the numerical aperture is very high and the effective mode area is very small. Further, nanofibres are exceptionally strong; the tensile strength is in excess of 5 GPa, which allows bending with radii of just a few microns. The high numerical aperture keeps the losses low, even for such tight bending. Light guided in a nanofibre can experience strong nonlinearities due to the small effective mode area and is associated with significant evanescent fields just outside the fibre’s surface. For fibre diameters below about 0.6 μm, the mode radius of guided light increases as the fibre diameter is further decreased, essentially because the “guiding power” of a thinner fibre becomes weaker. Most of the optical power then propagates in the evanescent field outside the fibre, for instance, for a 300 nm diameter ONF, about half of the energy of 633 nm light is carried outside the fibre. ONFs are therefore very sensitive to their immediate environment. As such, they are excellent candidates for sensors and can transmit light at visible and near IR wavelengths.
ONF sensors: interferometers, resonators and spectroscopy
The sensing capabilities of ONFs are being studied by the well-known Mazur Group at Harvard University. Using a two-step drawing process, the group has fabricated long, freestanding silica fibres with diameters down to 50 nm that show atomic-level surface smoothness and excellent diameter uniformity. The lengths can be up to tens of millimeter, yielding an aspect ratio of >50,000. In recognition of the creation of the strong evanescent fields associated with ONFs, while at the same time maintaining the coherence of the light, the group argues that the technology has the potential to yield sensitive and highly miniaturised optical sensors for physical, chemical and biological applications. They have modelled a sensing concept based on ONFs and a Mach-Zehnder interferometer, where two silica nanofibres with the same diameters are assembled onto a Mach-Zehnder structure (Figure 2). Part of the upper arm is used for sensing and is exposed to the environment and part of the lower arm acts as a reference and is isolated from the measurand. After travelling through the active area where analyte particles are exposed to the evanescent fields, the signal guided along the sensing arm meets the reference where a highly sensitive interferometry technique is used to measure the light’s phase shift (ΔΦ). The group has calculated the sensitivity of the device. For a ΔΦ value of 1.2×10−5 π, the expected detection limit for the number of particles is about 200 for a 10 nm diameter particle with a refractive index of 1.35, or 4 for a 100 nm diameter particle with a refractive index of 1.50. If the sensitive/active length of the ONF is 50 μm and is used in a biosensor, the corresponding sensitivity will be 10−4 of a monolayer or 10 pg/cm2, assuming a detection limit of 50 protein molecules with molecular mass of 50 kDa. This is a much higher sensitivity than Mach-Zehnder interferometry-based biosensors reported elsewhere, which is generally around 100 pg/cm2. An ONF-based Mach-Zehnder structure has been developed at the Chinese Zhejiang University of Technology which has been used to determine differing concentrations of isopropyl alcohol (IPA). The varying IPA concetrations are associated with changes in the solution’s refractive index which causes a phase shift of the guided mode. The sensor was able to detect an index variation of ∼10−6.
Again recognising that a potential application of ONFs is interferometric sensing, research at the University of Bonn has demonstrated that a single ONF can act as a Mach-Zehnder interferometer (Figure 3). If appropriately designed, the down- and up-tapers of an ONF can act as beam splitters. In the down-taper, the input fundamental mode (LP01) is split into two lowest circular modes (LP01 and LP02) which propagate through the waist of the fibre with different effective refractive indices due to the modal dispersion, and therefore accumulate different phase shifts. In the up-taper they are recombined into the LP01 mode. If the two modes meet in the up-taper in-phase, the intensity of the output LP01 mode will be maximised due to constructive interference and if they meet out of phase, the output signal will be minimised, the energy being coupled to higher, non-guided modes and eventually lost. Thus, an ONF becomes the equivalent of a Mach-Zehnder interferometer with two arms combined in space.
Optical microresonators based on evanescent field coupling have recently attracted a great deal of attention as biological or chemical sensors, as they are small and have the potential for high sensitivity, high selectivity and low detection limits. Specifically, the large evanescent fields yield high sensitivity; high Q-factors yield low limits of detection; and the corresponding small resonant bandwidths yield good wavelength selectivity. Accordingly, ONFs are being studied in this context and an example is research by the Optoelectronics Research Centre at the University of Southampton in the UK. In recognition of the fact that tightly coiled nanofibres can be used for miniature resonators, the Centre has developed a sensor dubbed a “coated all-coupling nanowire microcoil resonator” (CANMR) (Figure 4). To fabricate the device an expendable rod was coated with a layer of a low-loss polymer such as Teflon and an ONF with a diameter of 400 nm was wrapped around the rod. The whole structure was coated with the same polymer and the rod was then removed. The space left by the rod acts as a fluidic channel to transport a liquid sample to the resonator. The embedded nanofibre has a considerable fraction of its mode propagating in the fluidic channel, thus any change in the analyte properties reflects in a change of the mode properties at the CANMR’s output. Assuming a continuous-wave input, a change of the analyte’s refractive index will lead to a change of the effective index of the propagating mode, thereby shifting the mode relative to the resonance and thus modifying the transmission spectrum. The sensor has been shown to exhibit a sensitivity of 700 nm/RIU (refractive index units) at a wavelength of 970 nm and a limit of detection of 10−6-10−7 RIU. This means that it is possible to recognise an analyte molecule in 10 million molecules of solvent which is more than ten times more sensitive than any similar sensor previously proposed.
Several ONF sensors exploiting evanescent wave interactions in non-resonant structures have been demonstrated and an example is work by a group at the Chinese State Key Lab of Modern Optical Instrumentation at Zhejiang University. Here, the sensing element comprises a 680 nm diameter ONF taper coated with an 80 nm thick layer of gelatin. It is powered by a laser diode, operating at a wavelength of 1,550 nm and when exposed to moisture, the change in refractive index of the gelatin layer causes a portion of the power from the guided mode to be converted into the radiation mode, resulting in a relative humidity-dependent loss. The sensor has been tested with relative humidities ranging from 9 to 94 per cent and showed a response time of ∼70 ms which is significantly faster than other types of sensors based on conventional optical fibres. A similar concept has been adopted in a nanofibre-based hydrogen sensor, developed at the CIO, the Mexican research centre for optics. The fibre was coated with an ultra-thin (4 nm) layer of palladium which reacts with the gas through a well-characterised mechanism, causing attenuation of the evanescent field. Excitation was again from a 1,550 nm laser diode, with a power of ∼1 mW. The sensor was tested with hydrogen concentrations up to 5 per cent and the limit of detection was estimated to be 0.05 per cent. The response was reversible and the response time (T90) at 3.9 per cent hydrogen was ∼10 s, which is between 3 to 5 times faster than that of other optical hydrogen sensors so far reported.
Research at the Johannes Gutenberg University in Mainz is studying the role of ONFs in high sensitivity, surface fluorescence spectroscopy. The ONF was produced by stretching a standard, single-mode optical fibre while heating it with a hydrogen/oxygen flame, which led to fibres with a homogeneous waist diameter down to 100 nm. Using ONFs with a diameter of 320 nm and a length of 1 mm, molecules of a test compound 3,4,9,10-perylene-tetracarboxylic dianhydride (PTCDA) were adsorbed on the surface of the nanofibre and excited by 405 nm light from a diode laser with a power of ∼8 μW. The resulting fluorescence was captured and analysed and it was found that PTCDA surface coverages of as little as 1 per cent still gave rise to fluorescence spectra with a good signal to noise ratio and a sensitivity which exceeds those of previous studies by two orders of magnitude. Absorption from a white light (tungsten) source has also been studied and a schematic of the system used in this work is shown in Figure 5. Spectroscopy systems based on ONFs are being studied at the University of Tokyo. Here, a simple and compact ONF spectrometer has been developed which again relies on the evanescent field interacting with an analyte, in this case acetylene gas. The frequency resolution was comparable to that of the portable wavelength standard developed by NIST, the US National Institute of Standards and Technology, and the group argues that the spectrometer could act as a practical wavelength reference in the optical fibre communication band.
Tapered optical fibre nanosensors
Reducing the diameter of the tip of an optical fibre to sub-wavelength dimensions leads to a phenomenon similar to that discussed above, i.e. the creation of an evanescent field which produces a highly localised excitation source termed “near-field excitation”. The use of this effect in sensors arose from earlier research in the 1980s into near-field optical scanning microscopy, a technique which breaks the classical Abbé diffraction limit, allowing images of a sample to be obtained with nanometer-scale spatial resolution, typically 10-50 nm. The tapered tips are produced by a range of techniques which include focused ion beam (FIB) milling and chemical etching and Figure 6 shows a nanofibre tip with an aperture of 200 nm, produced by the FIB process.
This technology is more mature than the previously described class of ONF sensors and the first tapered optical fibre nanosensor (TOFN) was reported in Science back in 1992 by a group from the University of Michigan and was used for the measurement of pH. It involved an optical fibre with a tip tapered to between 100 and 1,000 nm with the walls coated with an aluminium film which ensured that total internal reflection occurred over the tapered region of the fibre. The tip was coated with a pH-sensitive dye and the sensor showed a remarkably rapid response, around 300 ms, which is approximately 100 times faster than that of more conventional fiber optic chemical sensors. Since then, several chemically responsive TOFNs have been reported. For example, workers from the Forschungszentrum Karlsruhe GmbH and the University of Regensburg developed sensors for chloride ions and pH which were based on 50 and 300 nm fibre tips coated with luminescent indicators whereby the analyte concentration was determined by measuring the luminescent decay time by a phase-modulation technique. The pH sensors used the indicator bromothymol blue and changes in the decay time were induced by resonance energy transfer from a ruthenium complex (the donor) to the bromothymol blue (the acceptor). The chloride sensor employed the chloride-carrier tridodecylmethylammonium chloride in a plasticised PVC membrane. Interest in TOFNs for chemical sensing applications has waned somewhat in recent years, most probably because they employ very similar principles (fluorescence, luminescence) and active materials (ion-responsive fluorophores and chromophores) as their macro-scale counterparts and other than being smaller, offer few advantages. As noted elsewhere in this issue, fibre optic chemical sensors have so far exerted little commercial impact. However, the biosensing applications of TOFNs are attracting growing interest, largely due to their high sensitivity and minute size which confers the unique ability to probe the volumes of individual cells.
Biosensing and sensing within single cells
In 1996, the first biosensor based on TOF technology was reported. This used a multimode optical fiber, tapered down to 40 nm at the tip. A layer of silver was applied to the sides of the fibre and antibodies which bind to the DNA adduct benzo[a]pyrene tetrol (BPT) were attached to the tip. It was found that the antibodies retained more than 95 per cent of their binding affinity for BPT and the BPT detection limit was approximately 300 zmol (300×10−21 mol). This is an important achievement, as BPT is a biomarker of human exposure to the carcinogen benzo[a]pyrene, a polycyclic aromatic hydrocarbon of great environmental and toxicological interest because of its mutagenic/carcinogenic properties and ubiquitous presence in the environment.
These encouraging, early results have led to significant research effort and recent years have seen the development of TOFNs for a range of quantities of biological importance, principally aimed at research in the clinical and healthcare sectors. The sensors are very similar to their predecessors and comprise tapered fibres with nanoscale diameter tips, coated with thin films of an optically opaque metal such as aluminium, silver or gold, with chemical or biological receptors such as antibodies, peptides or nucleic acids immobilised on the tip. Laser light is transmitted into the fiber, producing an evanescent field at the tip that excites target molecules bound to the bioreceptor molecules. The fluorescence originating from the analyte molecules is quantified by a conventional optical detection system. The minimal size of TOFNs has allowed the in vivo monitoring of biologically important compounds to be made within individual cells, even living cells. One of the leading TOFN research groups is at Duke University and by using an antibody as the bioreceptor the group has developed a TOFN which has been used to quantify the levels of BPT within rat liver epithelial cells. Following sensor calibration using differing BPT concentrations (Figure 7), a value of (9.6±0.2)×10−11 M was determined. Excitation was from the 325 nm line of a HeCd laser, focused onto a 400 μm delivery fibre to the ∼40 nm fibre tip. A schematic of the measuring system is shown in Figure 8. The group has developed similar sensors to determine the enzymatic activity of caspase-9, a marker for apoptosis (cell death). The sensor used a modified immunochemical assay format of a non-fluorescent enzyme substrate, leucine-glutamic acid-histidine-aspartic acid-7-amino-4-methyl coumarin (LEHD-AMC) and was used to probe a living MCF-7 cell (a breast cancer cell line isolated in 1970). The LEHD-AMC was cleaved during apoptosis by caspase-9, thus generating free AMC molecules that become fluorescent on laser excitation. The evanescent field was used to excite cleaved AMC and the resulting fluorescence signal was detected and quantified. By monitoring the changes in fluorescence, caspase-9 activity within a single, living cell was detected.
The growing body of research into sensors based on ONFs has shown that the technology has the potential to yield families of devices with a number of potentially beneficial features, notably rapid response times and high sensitivities. However, as noted in this issue’s Viewpoint, commercial success will only arise if these features can satisfy very specific user-requirements and this has yet to be demonstrated. Equally, the prospects for chemically responsive TOFNs remains uncertain. However, the unique ability of TOFNs to probe individual cells is leading to a new generation of tools that can detect the earliest signs of disease at the single cell level. They are expected to play a vital role and have the potential to change our fundamental understanding of many aspects of the life process itself.
Zhang, Y., Dhawan, A. and Vo-Dinh, T. (2011), “Design and fabrication of fiber-optic nanoprobes for optical sensing”, Nanoscale Research Letters, Vol. 6, p. 18