## Abstract

We analyzed the effect of fiber bending on spectral position and strength of the surface plasmon resonance arising due to the interaction of the fundamental mode guided in the core of the microstructured fiber with a metal layer. Fully vectorial simulations were performed using the finite element method with perfectly matched layers boundary conditions. To conduct the simulations, we adopted the concept of an equivalent bent fiber developed recently on the ground of transformation optics formalism. In this approach, the bent fiber with a metal layer is replaced by an equivalent fiber with appropriate spatial distributions of electric permittivity and magnetic permeability tensors. The obtained results explain the mechanisms responsible for the change in the SPR spectrum induced by bending and by the geometry of the microstructured fiber. By modifying the holes layout in the microstructured cladding, we designed the fiber, in which the depth of the surface plasmon resonance is in a high degree tunable by bending.

© 2013 OSA

## 1. Introduction

Surface plasmon resonance (SPR) based sensors are frequently used to detect various chemical and biological analytes changing a refractive index as a result of molecular interactions [1–5]. Subtle variations of the refractive index affect the resonance between the high-loss surface plasmon wave excited at the metal-dielectric interface and the incident wave. Excitation of the surface plasmon by a TM-polarized incident beam with a matched wave vector results in lower reflectivity. By measuring light reflectance, the resonance position related to the analyte refractive index is determined either as a function of wavelength at a fixed angle of incidence (spectral interrogation) or as a function of incidence angle for fixed wavelength (angular interrogation).

Conventional SPR sensors employing angular or spectral interrogation are based on Kretschmann’s configuration [6], in which a prism is used to excite the surface plasmon. Such devices are relatively complex and require fine tuning, therefore they can operate only in laboratory environment. Fiber-optic SPR sensors [7–20] overcome these limitations thanks to a simple design, high degree of integration, miniaturization and possibility of remote sensing even in harsh environment. A low cost and potentially broad applicability make them nowadays the promising class of the SPR sensors. The fiber-optic SPR sensors are limited only to wavelength interrogation but thanks to carefully designed sensing probes their performance can be competitive with conventional SPR sensors.

The simplest fiber-optic SPR sensors are based on a step index fiber with an exposed core coated with a metal layer [7,8] to allow for coupling between the guided modes and the surface plasmons. A performance of the sensing probes with removed cladding may be further increased by fiber tapering [9,10] or bending [11,12]. Theoretical studies on the U-shaped SPR probe [11] based on ray optics formalism, applicable only to multi-mode fibers of simple geometry, show that the sensitivity of the fiber-optic SPR probe increases with bending curvature. The increase in the sensitivity was accompanied by significant broadening of the resonance curve and the change in the resonance wavelength. The bend-induced change in the characteristics of the hetero-core SPR sensor was studied experimentally in [12]. It was shown that an increase in the curvature of the sensing probe amplifies the SPR loss and shifts the resonance wavelength.

In this work we numerically analyzed the effect of bending on SPR resonance in single-mode microstructured fibers. It has been already shown that microstructured fiber-optic SPR probes [14–20] containing an integrated metal-clad analyte channel allow for flexible shaping of the resonance characteristics. We show that much higher degree of bent-induced tunability of the SPR can be obtained in microstructured fibers than in conventional fibers reported in [11,12]. This effect is achieved by facilitating the power flow of the fundamental mode towards the metal layer by enlarging the pitch distance and reducing the size of selected air holes in the microstructured cladding. As the simplified ray-optics method proposed in [11] can’t be used to analyze the effect of bending in microstructured fibers with metal inclusions, we applied a rigorous numerical approach based on the finite element method (FEM). It employs the concept of the equivalent straight fiber with electric permittivity and magnetic permeability tensors derived in [21] using the transformation optics formalism [22].

## 2. Numerical method

Numerical modeling of straight microstructured fibers [23,24], including the SPR probes [14,16], is usually conducted using FEM, which allows to analyze waveguides with arbitrary geometry and boundary conditions. In principle, a 3-D FEM can be used to model directly the properties of bent waveguides but its high memory consumption strongly encourages the usage of two dimensional models based on the concept of an equivalent straight fiber [21,25], in which the change in the fiber geometry due to bending is transformed into appropriate distribution of material properties.

The most frequently used method of this type was proposed by Heiblum *et al.* in [25]. The authors showed that conformal coordinate transformation applied to the scalar Helmholtz equation is equivalent to the isotropic change in the refractive index in Cartesian coordinates. In this method, the fiber curvature is represented as an exponential change of the equivalent refractive index in the bending plane. Despite its usefulness, this method does not describe bend-induced transformation of permittivity and permeability and therefore cannot be used to model bent metal layers, which are key elements of the SPR sensors. Furthermore, the method is accurate only for the scalar Helmholtz equation, which is an approximation of the vectorial wave equation valid only for waveguides with low contrast of the refractive index. To model the effect of surface plasmon resonance in bent fibers, we adopted a recently proposed rigorous 2D representation of the curved waveguides [21], which overcomes the limitations of the Heiblum approach by use of the transformation optics formalism. In this approach, the transformed ε and µ tensors in the equivalent fiber are derived from Maxwell’s equations written in arbitrary coordinates.

In [22] Schurig *et al.* showed that the transformation between the coordinate systems x = {x^{1},x^{2},x^{3}} and x’ = {x^{1’},x^{2’},x^{3′}} is equivalent to the following change of the permittivity ε and permeability µ tensors:

Equations (1) and (2) can be used to determine the material properties in the equivalent straight fiber. In our model we used the transformation from Cartesian to cylindrical coordinate system proposed in [21], which does not deform the cross-section of the waveguide and conserves its length measured along its symmetry axis:

For the proposed transformation of the coordinate systems, the matrix ${\Lambda}^{i\text{'}}{}_{i}$takes the following form:

Transformation rules established by Eq. (7) and (8) can be used to model the bent waveguides made of any materials, including conducting and anisotropic media. We therefore adopted this approach to represent bent metal layers in the SPR sensors and bent perfectly matched layers (PML) [21] that are used to model bending loss.

## 3. Microstructured fiber design

Propagation of guided modes in the bent fiber of low aperturedepends mainly on equivalent values of ε_{eq} and µ_{eq} in XY plane, which increase linearly towards the bent top according to the following relation:

Basing on this intuitive understanding, we proposed three variants of the microstructured fiber showed in Fig. 3, in which the SPR effect is highly tunable by bending.

The fibers are based on the hexagonal lattice with the pitch distance of Λ = 8µm and the outside diameter equal to 125 μm, Fig. 4. A solid core is surrounded by air holes of the diameter d_{1} = 0.6Λ. The analyte channel of the diameter d_{c} = 4Λ located at 4Λ from the center of the fiber core is coated with a 40 nm gold layer. The diameter of three holes located between the core and the analyte channel is reduced to d_{2} = 0.2Λ to increase a bent-induced leakage of the fundamental mode towards the gold layer.

The fiber no. 1 has the simplest structure with the core surrounded only by two rings of air holes. The fiber no. 2 has an open analyte channel, which allows for measurements of the substances located outside of the fiber as well as for local deposition of the gold layer in the sensing region by external vacuum evaporation. In the fiber no. 3, a number of air hole rings was increased to reduce propagation and bending loss. The proposed fibers are tunable by bending radii of the order of few centimeters, facilitate the flow of the analyte due to large size of the analyte channel and are relatively easy tofabricate.

## 4. Simulation results

To calculate the complex propagation constants of the guided modes, we used the finite element method with perfectly matched layers boundary conditions. Wavelength dependence of the refractive index of silica glass was determined using the following Sellmeier’s equation [26]:

_{1}= 0.6961663, A

_{2}= 0.4079426, A

_{3}= 0.8974794, Z

_{1}= 0.0684043, Z

_{2}= 0.1162414 and Z

_{3}= 9.896161. Complex permittivity of the gold layer was approximated by the Drude model [27]:

^{∞}= 9.75, ω

_{p}= 1.36x10

^{16}[rad/s] and ω

_{τ}= 1.45x10

^{14}[rad/s].

The calculations were performed for the half of the fiber structure taking into account its plane symmetry. In the next paragraphs, we present the SPR spectra calculated for the mode with the electric field parallel to the bending plane, which is coupled more efficiently to the surface plasmons. In Figs. 5, 6, and 7 we show the results of simulations only for the fiber no.1, because the resonance characteristics for the other fibers are nearly identical as the differences in fibers’ microstructure are located far away from the region in which the surface plasmon resonance occurs. The comparison of the SPR spectra calculated for all the fibers is shown in Fig. 8. Discrepancies between the SPR spectra in the three fibers rise with the bending curvature but are practically negligible even for the lowest considered bending radius R = 5 mm.

As it is shown in Fig. 5, the straight fiber has a single, well defined plasmon resonance in the visible range. The resonance occurs at λ_{peak} = 588 nm for the analyte of the refractive index n_{a} = 1.33 and shifts by 14 nm to λ_{peak} = 602 nm for n_{a} = 1.34.

The resonance wavelength shift observed in response to the change in the refractive index of the analyte determines the sensitivity${S}_{\lambda}$of the sensing probe:

^{−5}RIU assuming that the position of the resonance wavelength is determined with the resolution of 0.1 nm. This figure is of the same order as the resolution of other microstructured SPR sensors reported earlier in literature [13–16].

Our simulations show that the bent-induced amplification of the resonance loss in the analyzed microstructured fibers is much higher than in the conventional fiber of a comparable diameter reported in [12]. In particular, for the bending radius R = 10 mm, the SPR loss is enhanced only by the factor of 2.5 in [12] compared to the factor of 200 in case of the analyzed fiber. For the considered microstructured fiber, the SPR spectrum is amplified 20 times for much larger bending radius R = 50 mm and up to 700 times for R = 5 mm, which is assumed to be the smallest bent radius the fiber can sustain without breaking. The fiber sensitivity does not change considerably upon bending. The resonance wavelength shifts by about 6 nm for the bending radius R = 5 mm but the shape of the normalized resonance curve does not broaden as it is in [11]. This is because the considered fiber is single mode and only the fundamental mode is coupled to the plasmon, Fig. 6.

The resonance wavelength shift Δλ calculated using FEM agrees relatively well with intuitive prediction based on Eq. (7), in which the wavelength shift Δλ is attributed solely to the bending-induced variation of the equivalent refractive index of the analyte:

As it is shown in Fig. 7, the field distribution of the fundamental mode changes significantly due to bending. Transition loss Γ between the fiber segments of different bending radii (R_{1},R_{2}) can be calculated using the following relation:

Abrupt changes in the fiber curvature leads to high transition loss, which, however, can be significantly reduced by a gradual increase in the bending radius. As it is shown in Table 1, the transition loss between the straight and the bent fiber with R = 5 mm is 13.2 dB, but if the additional segment of R = 50 mm is formed between them, it drops to 5.5 dB. The use of three additional segments of R = 50 mm, R = 25 mm and R = 10 mm decreases it further to 3.2 dB.

The confinement loss in the fiber without the metal layer can be greatly reduced by use of more air hole rings. The confinement loss in the straight fiber no.1 (with only two air hole rings) is about 2x10^{−2} dB/m in the resonance wavelength range, whereas for the fiber no. 3 (four air hole rings) it is reduced to about 4x10^{−9} dB/m. For the bent fiber a greater loss occurs because the fundamental mode leaks out of the microstructured cladding through the region with smaller air holes, Fig. 7. In this case, for R = 25 mm the loss rises to approximately 0.75 dB/m in the fiber no.1 and to 4x10^{−3} dB/m in the fiber no.3. Bending the fiber in the same plane but in the opposite direction (in this case R is denoted as negative) results in the decrease in bending loss because the fundamental mode is moved towards the holes with a larger filling factor. As it is shown in Table 2, for large bending radii, the loss can be reduced in both fibers below the level of the straight fiber, i.e., nearly two orders of magnitude for R = −50 mm.

In Fig. 8, we show the SPR spectra calculated for three variants of the proposed fiber design for the bending radius R = 5 mm. The differences are small enough to treat the results obtained for the fiber no.1 as universal, even in case of the small bending radii.

## 5. Summary

To analyze the effect of bending on the SPR spectra, we adopted a recently proposed rigorous approach, which employs the concept of equivalent fiber with appropriate spatial distribution of electric permittivity and magnetic permeability tensors derived on the ground of transformation optics formalism. We proposed and analyzed a design of the microstructured fiber, in which bending leads to strong amplification of the SPR spectrum caused by leaking of the fundamental mode towards the metal layer.The mode leakage is enhanced by increasing the pitch distance and by reducing the size of selected air holes in the region between the core and the analyte channel. As a result, the SPR resonance arising in the visible range is amplified by the factor of up to 700 for bending radius R = 5 mm. The shape of the resonance curve does not change upon bending, while the resonance wavelength raises by about 6 nm for R = 5 mm as a result of equivalent change in the refractive index of the analyte. The obtained sensitivity of 1400 [nm/RIU] is of the same order as reported earlier in [13–16] for other microstructured probes.

The proposed fiber designs can be used as integrated SPR sensors with an internal analyte channel (fiber no.1) or open structure (fiber no.2) capable of sensing *in situ* and operating on various length scales. Additionally, the fiber no. 2 can benefit from the possibility of local deposition of the metal layer by external vacuum evaporation, which will reduce the propagation losses in the fiber section without metal layer and allow for remote sensing using a single fiber. The simulations conducted for the fiber no.3 show that the propagation loss and bending loss in the leading sections of the microstructured fiber without a metal layer can be significantly reduced by adding additional rings of holes, which does not affect the SPR characteristics of the sensing probe.

## Acknowledgment

This work was supported by Wroclaw Research Center EIT + Ltd. in the frame of the NanoMat project “Application of Nanotechnology in Advanced Materials”, within the European Funds for Regional Development, POIG, Sub-action 1.1.2.

## References and links

**1. **J. N. Anker, W. P. Hall, O. Lyandres, N. C. Shah, J. Zhao, and R. P. Van Duyne, “Biosensing with plasmonic nanosensors,” Nat. Mater. **7**(6), 442–453 (2008). [CrossRef] [PubMed]

**2. **J. Homola, S. S. Yee, and G. Gauglitz, “Surface plasmon resonance sensors: review,” Sens. Actuat. Biol. Chem. **54**, 3–15 (1999).

**3. **J. Homola, “Present and future of surface plasmon resonance biosensors,” Anal. Bioanal. Chem. **377**(3), 528–539 (2003). [CrossRef] [PubMed]

**4. **E. Hutter and J. H. Fendler, “Exploitation of localized surface plasmon resonance,” Adv. Mater. **16**(19), 1685–1706 (2004). [CrossRef]

**5. **U. Jönsson, L. Fägerstam, B. Ivarsson, B. Johnsson, R. Karlsson, K. Lundh, S. Löfås, B. Persson, H. Roos, I. Rönnberg, S. Sjolander, E. Stenberg, R. Stahlberg, C. Urbaniczky, H. Ostlin, and M. Malmqvist, “Real-time biospecific interaction analysis using surface plasmon resonance and a sensor chip technology,” Biotechniques **11**(5), 620–627 (1991). [PubMed]

**6. **E. Kretschmann, “Determination of optical Constants of metals by excitation of surface plasmons,” Z. Phys. **241**(4), 313–324 (1971). [CrossRef]

**7. **R. Slavik, J. Homola, and J. Ctyroky, “Miniaturization of fiber optic surface plasmon resonance sensor,” Sens. Actuat,” Biol. Chem. **51**, 311–315 (1998).

**8. **J. van Gent, P. V. Lambeck, H. J. M. Kreuwel, G. J. Gerritsma, E. J. R. Sudhölter, D. N. Reinhoudt, and T. J. A. Popma, “Optimization of a Chemooptical Surface Plasmon Resonance Based Sensor,” Appl. Opt. **29**(19), 2843–2849 (1990). [CrossRef] [PubMed]

**9. **A. Diez, M. V. Andres, and J. L. Cruz, “In-line fiber-optic sensors based on the excitation of surface plasma modes in metal-coated tapered fibers,” Sens. Actuat. Biol. Chem. **73**, 95–99 (2001).

**10. **Y. C. Kim, W. Peng, S. Banerji, and K. S. Booksh, “Tapered fiber optic surface plasmon resonance sensor for analyses of vapor and liquid phases,” Opt. Lett. **30**(17), 2218–2220 (2005). [CrossRef] [PubMed]

**11. **R. K. Verma and B. D. Gupta, “Theoretical modelling of a bi-dimensional U-shaped surface plasmon resonance based fibre optic sensor for sensitivity enhancement,” J. Phys. D **41**(9), 095106 (2008). [CrossRef]

**12. **K. Takagi, H. Sasaki, A. Seki, and K. Watanabe, “Surface plasmon resonances of a curved hetero-core optical fiber sensor,” Sens. Actuat. A. **161**, 1–5 (2010).

**13. **M. Erdmanis, D. Viegas, M. Hautakorpi, S. Novotny, J. L. Santos, and H. Ludvigsen, “Comprehensive numerical analysis of a surface-plasmon-resonance sensor based on an H-shaped optical fiber,” Opt. Express **19**(15), 13980–13988 (2011). [CrossRef] [PubMed]

**14. **M. Hautakorpi, M. Mattinen, and H. Ludvigsen, “Surface-plasmon-resonance sensor based on three-hole microstructured optical fiber,” Opt. Express **16**(12), 8427–8432 (2008). [CrossRef] [PubMed]

**15. **A. Wang, A. Docherty, B. T. Kuhlmey, F. M. Cox, and M. C. J. Large, “Side-hole fiber sensor based on surface plasmon resonance,” Opt. Lett. **34**(24), 3890–3892 (2009). [CrossRef] [PubMed]

**16. **A. Hassani and M. Skorobogatiy, “Design criteria for microstructured-optical-fiber-based surface-plasmon-resonance sensors,” J. Opt. Soc. Am. B **24**(6), 1423 (2007). [CrossRef]

**17. **A. Hassani and M. Skorobogatiy, “Photonic crystal fiber-based plasmonic sensors for the detection of biolayer thickness,” J. Opt. Soc. Am. B **26**(8), 1550–1557 (2009). [CrossRef]

**18. **A. Hassani and M. Skorobogatiy, “Surface Plasmon Resonance-like integrated sensor at terahertz frequencies for gaseous analytes,” Opt. Express **16**(25), 20206–20214 (2008). [CrossRef] [PubMed]

**19. **M. Skorobogatiy and A. V. Kabashin, “Photon crystal waveguide-based surface plasmon resonance biosensor,” Appl. Phys. Lett. **89**(14), 143518 (2006). [CrossRef]

**20. **B. Gauvreau, A. Hassani, M. Fassi Fehri, A. Kabashin, and M. A. Skorobogatiy, “Photonic bandgap fiber-based surface plasmon resonance sensors,” Opt. Express **15**(18), 11413–11426 (2007). [CrossRef] [PubMed]

**21. **D. M. Shyroki, “Exact equivalent straight waveguide model for bent and twisted waveguides,” IEEE Trans. Microw. Theory Tech. **56**(2), 414–419 (2008). [CrossRef]

**22. **D. Schurig, J. B. Pendry, and D. R. Smith, “Calculation of material properties and ray tracing in transformation media,” Opt. Express **14**(21), 9794–9804 (2006). [CrossRef] [PubMed]

**23. **J. Olszewski, M. Szpulak, and W. Urbańczyk, “Effect of coupling between fundamental and cladding modes on bending losses in photonic crystal fibers,” Opt. Express **13**(16), 6015–6022 (2005). [CrossRef] [PubMed]

**24. **G. Statkiewicz-Barabach, J. Olszewski, M. Napiorkowski, G. Golojuch, T. Martynkien, K. Tarnowski, W. Urbanczyk, J. Wojcik, P. Mergo, M. Makara, T. Nasilowski, F. Berghmans, and H. Thienpont, “Polarizing photonic crystal fiber with low index inclusion in the core,” J. Opt. **12**(7), 075402 (2010). [CrossRef]

**25. **M. Heiblum and J. Harris, “Analysis of Curved Optical-Waveguides by Conformal Transformation,” IEEE J. Quantum Electron. **11**(2), 75–83 (1975). [CrossRef]

**26. **M. Bass, *Handbook of Optics, 3rd edition, Vol. 4: Optical Properties of Materials, Nonlinear Optics, Quantum Optics* (McGraw-Hill, 2009).

**27. **K. Kurihara, K. Nakamura, E. Hirayama, and K. Suzuki, “An absorption-based surface plasmon resonance sensor applied to sodium ion sensing based on an ion-selective optode membrane,” Anal. Chem. **74**(24), 6323–6333 (2002). [CrossRef] [PubMed]