546
OPTICS LETTERS / Vol. 40, No. 4 / February 15, 2015
Plasmonic nanoshell functionalized etched fiber Bragg gratings for highly sensitive refractive index measurements Jörg Burgmeier,1,2,* Amin Feizpour,2 Wolfgang Schade,1,3 and Björn M. Reinhard2 1
Department of Fiber Optical Sensor Systems, Fraunhofer Heinrich-Hertz-Institute, EnergieCampus, Am Stollen 19B, 38640 Goslar, Germany 2
Department of Chemistry and The Photonics Center, Boston University, 8 St. Mary’s St., Boston, Massachusetts 02215, USA 3
Department of Applied Photonics, Institute of Energy Research and Physical Technologies, EnergieCampus, Am Stollen 19B, 38640 Goslar, Germany *Corresponding author:
[email protected] Received December 4, 2014; revised January 9, 2015; accepted January 10, 2015; posted January 14, 2015 (Doc. ID 229050); published February 9, 2015
A novel fiber optical refractive index sensor based on gold nanoshells immobilized on the surface of an etched single-mode fiber including a Bragg grating is demonstrated. The nanoparticle coating induces refractive index dependent waveguide losses, because of the variation of the evanescently guided part of the light. Hence the amplitude of the Bragg reflection is highly sensitive to refractive index changes of the surrounding medium. The nanoshell functionalized fiber optical refractive index sensor works in reflectance mode, is suitable for chemical and biochemical sensing, and shows an intensity dependency of 4400% per refractive index unit in the refractive index range between 1.333 and 1.346. Furthermore, the physical length of the sensor is smaller than 3 mm with a diameter of 6 μm, and therefore offers the possibility of a localized refractive index measurement. © 2015 Optical Society of America OCIS codes: (060.3735) Fiber Bragg gratings; (060.2370) Fiber optics sensors; (060.2280) Fiber design and fabrication; (220.4241) Nanostructure fabrication; (310.6628) Subwavelength structures, nanostructures. http://dx.doi.org/10.1364/OL.40.000546
Fiber optical sensors are attracting more and more attention because of a series of unique properties, including small size, high chemical stability, multiplexing capability, low weight and insensitivity against electromagnetic fields [1]. Fiber Bragg grating (FBG) sensors especially were demonstrated to have excellent abilities for a variety of applications, like strain [2], temperature [2], pressure [3], and refractive index sensing [4], because of their robustness, easy fabrication, and sensing principle in reflection mode. Refractive index measurements utilizing fiber grating sensors have been demonstrated with different approaches in the past, e.g., with etched FBGs [4], long period gratings [5], and tilted Bragg gratings [6]. To achieve an enhancement in the sensitivity for surrounding refractive index (SRI) measurements by exciting surface plasmon resonances (SPR), noble metal coatings [7] and, more recently noble metal nanoparticles [8] have been applied to the surface of optical fibers. These sensor types offer a great potential for a variety of applications, because of the combination of the small size and flexibility of fiber sensors with the excellent refractive index sensitivity of the SPR. However, each sensing approach has its individual strengths and weaknesses because of the used type of fiber (single mode/multi-mode), noble metal coating (layer/particle), measurement mode (transmission/reflection) and measured physical value (wavelength/intensity). Here we introduce an intensity-based sensing approach with a fiber Bragg grating (FBG) that works in reflection mode, resulting in a simple and cost-efficient setup. In our novel sensing scheme, an etched optical single-mode fiber, including an FBG, is coated with gold nanoshell particles to facilitate highly sensitive measurements of the SRI. We used a single-mode fiber to achieve 0146-9592/15/040546-04$15.00/0
a well-defined Bragg reflection from the FBG and to avoid perturbations associated with different mode distributions. Although FBGs have already been combined with plasmonic systems in the past, the grating was generally used to couple light from the fundamental mode into cladding modes, and therefore requires a signal analysis in transmission [9]. Furthermore, in most cases, propagating surface plasmon polariton sustaining films of gold were used, which leads to a more complex fabrication of homogenous layers, because of the curved surface of the fiber. The gold nanoshell based coating used in our experiments utilizes the excitation of localized surface plasmons (LSPs) in individual particles and clusters. Nanoparticle coated fibers are available via an uncomplicated surface modification step and have the additional advantage that clusters of nanoparticles can localize electromagnetic hot spots in the gaps between the nanoparticles, where the local electric field is enhanced by many orders of magnitudes. Figure 1 shows the experimental setup used for the fiber preparation and characterization. A fiber coupled broadband emitting a superluminescent diode (SLED) with a center wavelength of 1.55 μm, a 3 dB bandwidth
Fig. 1. Experimental setup used for sensor preparation and characterization. © 2015 Optical Society of America
February 15, 2015 / Vol. 40, No. 4 / OPTICS LETTERS
of 90 nm, and an output power of 2.5 mW was used as a light source. The polarization was controlled via a linear polarizer and a rotatable half-wave plate, installed in a fiber-to-fiber bench. The wavelength and the intensity of the reflected light from the FBG were analyzed by a spectrometer (NQ512, OceanOptics) with a resolution of 0.45 nm. A standard single-mode fiber complying with ITU-T Recommendation D and including an FBG at the fiber end was used. The FBGs had a grating length of 0.7 mm, a center wavelength of 1543 nm, a 3 dB bandwidth of 1.6 nm, and were inscribed with a point-by-point technique with femtosecond laser pulses. The used setup and inscription process is described elsewhere [10]. The main fiber preparation steps are shown in Fig. 2, including the chemical etching of the fiber’s glass coating material, the silanization of the fiber core to induce a positive surface charge, and the nanoparticle binding. A removal of the glass cladding is necessary to allow an evanescent interaction of the light guided in the fiber’s core with the immobilized nanoparticles and the surrounding medium. The fiber was therefore first etched by hydrofluoric acid (HF) with a concentration of 49% from a cladding diameter of 125 μm down to around 10 μm, while in a second etching process, an HF concentration of 4.9% was used to etch the fiber with a slower etching rate to the final diameter of approximately 6 μm. The total etching procedure takes about 1 h. The Bragg wavelength blue shifts as soon as the diameter is small enough for the guided electromagnetic field penetrating in the surrounding media. This is because of the fact that the Bragg wavelength of the FBG depends on the effective refractive index at the grating position, including the refractive index of the surrounding medium. Therefore, the reflected signal of the FBG was monitored during the etching to control the process. The higher concentrated HF was replaced by the lower concentration, as soon as the Bragg wavelength started to shift to a lower wavelength. The etching with the lower HF concentration has been continued until a total Bragg wavelength shift of around 8.1 nm was achieved, corresponding to a final
Fig. 2. Fiber preparation steps: (a) structure of the untreated single-mode fiber, (b) fiber tip after etching process, (c) silanization of glass surface, and (d) fiber tip with electrostatically bound nanoshell particles.
Fig. 3.
547
Reflecting spectra of the unetched and etched FBG.
fiber diameter of around 6 μm. This wavelength shift has been chosen as the standard for all measurements, because a further etching to smaller fiber diameters resulted in a significant decrease of the Bragg-reflected light intensity. In Fig. 3, the reflecting spectra of the FBG before and after the etching procedure are shown. The data points measured by the spectrometer are fitted by a Gaussian. After the etching process, the fiber was intensively rinsed with DI water and immersed in ethanol before the silanization. Because the used gold nanoshell particles have a negatively charged surface, an electrostatic binding strategy was chosen, where a positively charged fiber surface was generated by binding (3-Aminopropyl) trimethoxysilane (APTMS, 97% Sigma-Aldrich) to hydroxyl groups on the glass surface. Therefore, the fiber was immersed for 15 min in a freshly prepared solution of 5% v/v APTMS dissolved in ethanol, while the solution was continuously stirred. After the silanization process, the fiber was cleaned by incubating in fresh ethanol for 15 min to remove unbound APTMS, and dried for 1 h at room temperature. The gold nanoshell particles (Nanospectra Biosciences) used in our experiments were packaged in water and stored at 4°C. We chose nanoshells that consist of a gold nanoparticle layer around a dielectric (here SiO2 ) core for our experiments since their optical properties can be rationally tuned over a broad spectral range through the ratio of the gold layer thickness to the total diameter [11]. Individual nanoshells had a total diameter of around 155 nm and a peak resonance wavelength of approximately 800 nm, whereas the extinction spectrum of immobilized particles on a glass slide in Fig. 4(a) shows that the nanoparticle coating also provides
Fig. 4. (a) Extinction spectrum of nanoshells immobilized on a glass surface and (b) intensity and Bragg wavelength during nanoshell binding.
548
OPTICS LETTERS / Vol. 40, No. 4 / February 15, 2015
significant extinction at a wavelength of 1.55 μm. The broadened and red-shifted extinction spectrum of the nanoparticles used in this work is induced because of electromagnetic coupling between nanoparticles in spontaneously formed clusters on the glass surface [12,13]. In addition to clustering, an agglomeration of nanoparticles results in a broadening of the extinction spectrum. Furthermore, nanoshells are isotropic nanomaterials, which are extremely useful for an efficient interaction of the surface bound nanoparticles with the evanescent fields surrounding the fiber. Before immersing the fiber sample into the nanoshell solution, the particles were cleaned by centrifuging the solution with an acceleration of 0.4 g for 10 min and replacing the resolvent with fresh DI water. Immediately after the cleaning process of the particles, the silanized fiber was immersed in 1.5 mL of gold nanoshell solution with a concentration of around 1 × 109 particles per mL. While immersing the fiber sensor in the nanoparticle solution, nanoshells bind electrostatically to the silanized fiber surface. Because metal nanoparticles have a high polarizability and the nanoparticle extinction spectrum overlaps with the Bragg wavelength, the evanescently guided part of the light can efficiently interact with the nanoparticles on the fiber surface and excite LSPs. The excited LSPs can decay radioactively and dissipatively, resulting in losses for the guided light in the fiber’s core, and therefore in a decrease of the reflected signal from the FBG. The attenuation mechanisms of nanoparticles on a fiber surface for the evanescent part of light guided in a fiber core are described in the publication of Chau et al. [14]. Figure 4(b) shows the temporal development of the reflected light intensity and Bragg wavelength during the immersion of the fiber in the nanoshell solution. Because of the binding of nanoshells on the fiber core surface, the intensity of the Bragg reflection decreases, while the Bragg wavelength red-shifts, caused by the increasing effective refractive index. To ensure an acceptable signal to noise ratio after coating the fiber with nanoparticles, the fiber was removed from the particle solution once the reflected signal from the FBG had dropped to 3% of the initial intensity, corresponding to a typical Bragg wavelength shift of 200 pm. After the nanoparticle coating step, the fiber was rinsed with DI water to remove unbound nanoparticles. The achieved particle density on the fiber surface was estimated to be approximately 2.0% at the grating area and 0.5% elsewhere by analyzing SEM images (Fig. 5). The higher particle density at the grating area is induced by an increased binding affinity of the particles at the microstructure of the FBG, caused by the higher etching rate of the femtosecond laser modified glass volume [15]. These microstructures provide a natural template for the guided assembly of nanoparticle clusters embedded in the fiber, which facilitate nonlinear spectroscopy methods. To characterize the sensitivity for refractive index changes, the sensor was immersed in different mixtures of ethanol–water and the reflected signal from the FBG was measured with the setup shown in Fig. 1. The corresponding refractive index of the ethanol–water mixtures was calculated by the data published elsewhere [16]. Figure 6(a) shows the reflection spectra of the bare
Fig. 5. SEM image of etched FBG with nanoshell coating and increased particle density at the microstructured fiber grating area.
Fig. 6. Reflected spectra for different SRI: (a) bare fiber and (b) nanoshell coated fiber.
etched fiber sensor before the silanization for different SRIs. The amplitude of the Bragg reflection remains constant for SRI changes, while the Bragg wavelength shifts with a sensitivity of 13.3 nm per refractive index unit (RIU). In comparison, the spectral response of the nanoshell coated sensor is presented in Fig. 6(b). Thereby, the wavelength sensitivity remains nearly unchanged with 13.5 nm/RIU [Fig. 7(a)], whereas the intensity of the Bragg reflection shows a strong dependency on the SRI and decreases with increasing refractive index [Fig. 7(b)]. In the refractive index range between 1.333 and 1.346, the sensor response can be approximated by a linear function with a slope of −4400%∕RIU. The sensor response is not expected to be linear in a larger range. Furthermore, the range of measurement is limited to a maximum refractive index corresponding to the refractive index of the fiber’s core, which is typically around 1.45 for standard single mode fibers. The sensitive length of the fiber used for these experiments was 2.7 mm. The error bars shown in the data of Fig. 7(b) correspond to the standard deviation of five
Fig. 7. SRI dependent sensor response of bare and nanoshell coated fiber tip for (a) Bragg wavelength and (b) amplitude.
February 15, 2015 / Vol. 40, No. 4 / OPTICS LETTERS
Fig. 8. Response of the nanoshell coated fiber sensor for different SRI by immersing the sensor in different ethanol–water mixtures.
individual measurements. The refractive index sensitivity of the reflected intensity was also analyzed for differently polarized light by rotating the half-wave plate. Thereby, we observed no significant differences in the sensitivity for SRI changes. Furthermore, we analyzed the sensor response also for a fiber coating with 40 nm solid gold nanoparticles and gold nanorods with an LSP resonance at a wavelength of 530 nm and 1.45 μm, respectively. Both types of nanoparticles where immobilized on the fiber surface, but showed a much lower sensitivity for SRI changes, even for much higher particle densities on the fiber surface. This observation can be explained by the significant wavelength difference between the Bragg wavelength and the LSP resonance of the 40 nm solid gold particles, whereas for randomly aligned nanorods the anisotropic extinction characteristic prevents an efficient coupling between the evanescent field and the LSPs [17]. The sensitivity of the presented sensing principle could be further improved by using a monochromatic light source, emitting at a wavelength corresponding to the slope of the Bragg reflection in combination with a photo diode. Thereby, the intensity and wavelength refractive index dependency would contribute to the total refractive index sensitivity. Within this context, the reflecting profile could also be optimized, e.g., by using more complex grating structures like phase-shifted fiber Bragg gratings [18] to achieve a steeper slope and hence a further increasing of the refractive index sensitivity. Figure 8 shows the peak intensity of the Bragg reflection for an increasing and decreasing SRI respectively. Therefore, the sensor was immersed into vials with different water–ethanol mixtures for 3 min each. Between the individual immersions, the sensor was exposed to air.
549
A refractometer based on a nanoshell coated singlemode fiber, including a fiber Bragg grating, has been demonstrated. The sensor works in reflection mode, shows a reliable sensor response, and is easy to fabricate. The back reflected light intensity shows a high refractive index sensitivity of −4400%∕RIU. The sensitive length of the used fiber samples was approx. 2.7 mm with a diameter of 6 μm, and therefore allows for a localized refractive index measurement. Furthermore, the interrogation unit can be very cost-efficient because the sensing principle is based on an intensity measurement. A chemical or biological sensing could be achieved by crosslinking the fiber or particle surface with appropriate molecules or bioreceptors. This work was supported by the U.S. Department of Energy, Office of Basic Energy Sciences, Division of Materials Science and Engineering under Award DOE DE-SC0010679. References 1. F. T. S. Yu and S. Z. Yin, Fiber Optic Sensors (Marcel Dekker, 2002). 2. W. W. Morey, G. Meltz, and W. H. Glenn, Proc. SPIE 1169, 98 (1989). 3. M. G. Xu, L. Reekie, Y. T. Chow, and J. P. Dakin, Electron. Lett. 29, 398 (1993). 4. A. Iadicicco, A. Cusano, A. Cutolo, R. Bernini, and M. Giordano, IEEE Photon. Technol. Lett. 16, 1149 (2004). 5. V. Bhatia and A. Vengsarkar, Opt. Lett. 21, 692 (1996). 6. G. Laffont and P. Ferdinand, Meas. Sci. Technol. 12, 765 (2001). 7. R. C. Jorgenson and S. S. Yee, Sens. Actuators B 12, 213 (1993). 8. S.-F. Cheng and L.-K. Chau, Anal. Chem. 75, 16 (2003). 9. Y. Shevchenko and J. Albert, Opt. Lett. 32, 211 (2007). 10. J. Burgmeier, W. Schippers, N. Emde, P. Funken, and W. Schade, Appl. Opt. 50, 1868 (2011). 11. R. Averitt, S. Westcott, and N. Halas, J. Opt. Soc. Am. B 16, 1824 (1999). 12. T. Jensen, L. Kelly, A. Lazarides, G. C. Schatz, and J. Cluster, Science 10, 295 (1999). 13. B. Yan, S. V. Boriskina, and B. M. Reinhard, J. Phys. Chem. C 115, 24453 (2011). 14. L.-K. Chau, Y.-F. Lin, S.-F. Cheng, and T.-J. Lin, Sens. Actuators B 113, 100 (2006). 15. A. Marcinkevičius, S. Juodkazis, M. Watanabe, M. Miwa, S. Matsuo, H. Misawa, and J. Nishii, Opt. Lett. 26, 277 (2001). 16. T. A. Scott, Jr., J. Phys. Chem. 50, 406 (1946). 17. J.-M. Renoirt, M. Debliquy, J. Albert, A. Ianoul, and C. Caucheteur, J. Phys. Chem. C 118, 11035 (2014). 18. J. Burgmeier, C. Waltermann, G. Flachenecker, and W. Schade, Opt. Lett. 39, 540 (2014).
