Open Access
Add to BibSonomyAbstract
A novel differential phase detecting surface plasmon resonance (SPR) sensor based on white-light spectral interferometry is presented. Our proposed scheme employs a white-light source for SPR excitation and measures the corresponding SPR phase change at the optimized coupling wavelength with fixed angle of incidence across the visible spectrum. Compared to existing laser based phase detecting schemes, this system offers optimal sensitivity and extended dynamic range of measurement without any compromise in phase detection resolution. Results obtained from sodium chloride solutions indicate that the detection limit is 2.6 × 10−7 RIU over a refractive index range of 10−2 RIU, which is considerably wider than that achievable by existing laser based approach, thus making our scheme very attractive for practical SPR sensing applications.
©2011 Optical Society of America
1. Introduction
With decades of research and perfection, surface plasmon resonance (SPR) sensor has been indisputably demonstrated as a practical label-free sensing technology for numerous biology and chemistry applications [1]. In order to extract useful information from SPR systems, a number of detection techniques have been developed. These may be classified into four general categories by nature of operation, i.e. intensity, angular, wavelength, and phase. Theoretical detection limit of the first three categories was studied in a recent article and lies between 10−6 to 10−7 refractive index units (RIU) [2]. The introduction of laser-based phase detecting SPR systems with differential scheme [3–5] has further enhanced the detection limit to 10−8 and even 10−9 RIU. Although such high sensitivity might be sufficient for studying the interaction of bio-molecules, there are limitations as well as stringent prerequisites, namely (i) high detection sensitivity is usually restricted to a narrow dynamic range of measurement, i.e. typically 10−4 RIU, because of the highly nonlinear nature of SPR phase jump; (ii) typically the metal film thickness has to be within 1 to 2 nm from the optimal value [5], which makes it extremely difficult to provide high sample throughput. Consequently, the applicability of high detection sensitivity is rather limited as the refractive index change may easily exceed the linear region. Moreover, the phase performance suffers as the employed laser wavelength is non-optimized. To address the first limitation, temporal modulation technique which measures the higher harmonics has been proposed [6,7]. Expanding the dynamic range of the phase-sensitive region becomes essentially the next stage of SPR sensor development. However, not much effort has been given to address the wavelength issue. In our recent work, we reported a new SPR spectral phase scheme demonstrating which is practically possible to achieve biosensing with wide dynamic range and high detection resolution simultaneously [8]. In this paper, sensing performance of our spectral phase technique has been analyzed through simulation and experimental investigations. We show quantitatively that the calculated limit of detection is in the order of 10−7 RIU over a dynamic range of 10−2 RIU.
2. Theoretical investigation
The SPR effect on the electromagnetic wave can be analyzed by classical Fresnel equations [3,9]. The resultant wavelength dependent complex reflection coefficient of every p- and s- polarized spectral components can be expressed as
where Rp,s(λ) and φ p,s(λ) are the spectral reflectance and phase change of the attenuated total reflection (ATR) of corresponding polarization. A dip in the p-polarized spectrum can be observed at the optimized wavelength where maximum photon-to-surface plasmon transformation occurs while the s-polarized spectrum remains unaffected. By extracting the position of such spectral dip [10], or evaluating the accompanied differential phase shift Δφ(λ) = φ p(λ) – φ s(λ), the value of refractive index variations (ΔRI) can be effectively monitored. Since the frequency of the visible spectrum is in the order of 1015 Hz, the most practical way to determine Δφ(λ) is via spectral interferometry. The superposition of the white-light spectra from a common source but traversing different optical path contains a modulation term which is known as spectral interference [11]. In our case of SPR, the useful phase information is encoded in the spectral interferogram. Thereby, p- and s- polarized spectra interfere among themselves and the interferograms are obtained by using a polarizing beamsplitter. Spectral interference signals are then detected in corresponding polarization channels separately. Effectively, we have two independent spectral interferometers working simultaneously in parallel. The spectral signals recorded by the spectrometer with a Gaussian response function can be represented as [12]where Vp,s(λ) is a spectral visibility term related to Rp,s(λ), φ air is the phase introduced by the optical path difference between the two interferometric paths in air. Here the effect of dispersion is negligible because of the compensation effect introduced by a prism in the reference arm of the interferometer. The subscripts p, s, and noise carry their respective definitions. By mutual subtraction of the phase terms in Eq. (2), φ air and φ noise are effectively eliminated, and the spectral phase difference Δφ(λ) of the entire white-light spectrum can be extracted.Based on Fresnel equations and published data for the prism/metal/dielectric sensing structure [13], numerically modeling was performed with BK7 glass as the Kretschmann prism coupler, gold film thickness of 49nm, incident angle at 75.3°, and pure water in contact with the gold film. The corresponding ATR intensity and spectral phase profile of the p-polarization are shown in Fig. 1(a) . By introducing an optical path difference of 170µm between the two interference paths and assuming an optical resolution of spectrometer of 0.05nm full-width-half-maximum (FWHM), we generate the initial noise-free spectral interferograms based on Eq. (1) and (2) as shown in Fig. 1(b). Since SPR does not affect the s-polarization, it is included in the simulated interferogram as reference. Figure 1(b) demonstrates that the interference amplitude is significantly reduced at the optimal wavelength, i.e. about 645nm. In addition, the phase jump is designated by the shift of p-polarized interference fringes with respect to those of s-polarization as revealed in Fig. 1(b), i.e. changing from approximately in-phase before SPR to anti-phase afterwards.
Fig. 1 (a) (Color online) Spectral reflectance and phase of p-polarization obtained from Fresnel equations; 1(b) simulated zero mean spectral interferogram based on optical path difference of 140 microns and resonance data of Fig. 1(a).
The theoretical resolution of our white-light spectral interferometer can be estimated over a wide measurement range from 1.3330 to 1.3505 RIU, i.e. increment from distilled water to 10% sodium chloride (NaCl) solution by weight [14]. Adapting the linear approximation formula [2], the limit of detection (LOD) can be expressed as
where σ RI is the LOD, δn is the refractive index change of the bulk medium, δY is the calculated phase shift, and σ SD is the estimated phase noise. By taking a small ΔRI of 0.0017 from 1.3330 to 1.3347 RIU, i.e. replacing distilled water with 1.0% NaCl solution by weight, the spectral phase has shifted by about 82.76° at 645.30nm. Based on a phase measurement noise of 0.01° [3,5] with Eq. (3), our LOD is found to be 2.05 × 10−7 RIU. Using the same approach, we also calculated the LOD for up to 10.0% by weight of NaCl solutions. The spectral phase response of selected wavelengths with increment of NaCl concentration is shown in Fig. 2(a) . The LOD of our system from 1.3330 to 1.3505 RIU is summarized as Fig. 2(b) based on every optimal wavelength. From these figures, it is obvious that this system maintains high sensitivity over the entire measurement range by pursuing the optimal linear phase response across the visible spectrum.Fig. 2 (a) (Color online) Spectral phase response of selected wavelength with increment NaCl concentration; 2(b) estimated limit of detection of our white-light spectral interferometer based on simulated data from Fresnel equations [9] with 0.01° stability and reference [14].
3. Experimental configuration
3.1 Optical setup
A Michelson spectral interferometer [15] used in our experiments is shown in Fig. 3 . The light source is a 40W warm-white light emitting diode (LED) [16] with a central wavelength of 600nm and a FWHM bandwidth from 510nm to 690nm. The LED is fiber-coupled and directed by a collimator through a broadband linear polarizer oriented 45° to the optical axis. The broadband non-polarizing beam-splitter divides the incoming polarized LED spectrum into the reference and probe path respectively. Two identical BK7 Kretschmann SPR sensor prisms are inserted into the interferometric paths as coupling prisms, and for compensating the excessive dispersion introduced by the prisms themselves. The sensing surfaces are coated with gold film of 49nm nominal thickness and the incident angle is approximately 73°. Two high precision λ/20 mirrors are placed at the ends with one of them being mounted on a translation stage so that the optical path difference in air can be adjusted. The broadband polarizing beamsplitter at the exit of the interferometer divides the mixed interferogram into corresponding p- and s- polarizations. The two collimators collect the optical signal and transmit to the dual-channel grating spectrometer via the optical fibers. The FWHM optical resolution of our spectrometer is 0.05nm covering 600nm to 800nm for both channels. The spectral interferogram is adequately resolved and processed by the personal computer to extract the differential phase. Since our phase information is obtained from the spectral domain, the system does not require temporal modulation which is required for most existing SPR phase detection schemes. The Michelson interferometer is placed inside a thermally-shielded enclosure during operation. Effectively, our system consists of no moving parts or optoelectronic modulators, thereby greatly simplifying its operation.
Fig. 3 (Color online) Setup of our white-light SPR spectral interferometer for differential phase measurement: 1, 40W warm-white light emitting diode (LED); 2, collimator; 3, broadband linear polarizer oriented 45° to the optical axis; 4, broadband non-polarizing beamsplitter; 5, SPR probe cell which is attached to the peristaltic pump; 6 and 8, λ/20 high precision mirrors; 7, SPR reference cell which is filled with distilled water; 9, broadband polarizing beamsplitter which divides the two polarizations into different path; 10 and 11, collimators that collect the interference fringes into corresponding channel; 12, dual channel spectrometer; 13, personal computer for signal processing and data storage.
3.2 Verification of system performances
To demonstrate the features of both high phase sensitivity and wide dynamic range, various NaCl solution concentrations, i.e. 0%, 0.5%, 1%, 2%, 4%, 6%, 8%, and 10% by weight were injected into the probe cell. The spectral oscillations were first processed by Fourier zero-order filtering to remove the background intensity to obtain spectral signals of Eq. (2), as shown in Fig. 4 . The plots in Fig. 4(a) show the spectral oscillations of the p- and s- polarizations when the SPR sensor head was filled with distilled water. The existence of optimized SPR transformation was observed by (i) the spectral visibility of the optimized wavelength at 645nm in the p-polarized signal drops to a minimum; (ii) there is no sign of SPR excitation in the s-polarized signal; (iii) a significant spectral phase shift is observable between the p- and s-polarizations starting at the maximum attenuated wavelength. These observations are in good agreement with the simulation plot of Fig. 1(b). With the addition of 10% sodium chloride solution, i.e. 1.75 × 10−2 ΔRI, the optimized coupling shifts to a much longer wavelength, i.e. 665nm, yet these three distinctive SPR characters remain unchanged as shown in Fig. 4(b). These interferograms indeed show that this system is capable of detecting the differential phase change upon relatively large refractive index change.
Fig. 4 (Color online) Experimental spectral interference fringes acquired at (a) the presence of distilled water and (b) after injection of 10% NaCl solution into the probe cell, the dip in TM (p-polarized) spectral fringe amplitude and corresponding phase shift between TM and TE (s-polarized) fringes imply the optimized SPR occurs at about 646nm with water and it was shifted to about 675nm with 10% NaCl solution.
4. Results and discussion
In order to calculate the differential phase, windowed Fourier transform (WFT) [17] was adopted to extract φ p(λ) and φ s(λ) from respective p- and s- polarized spectral interferograms. The SPR induced differential phase jump Δφ(λ) of A: distilled water, B: 2%, C: 4%, D: 6%, E: 8% and F: 10% sodium chloride solutions were calculated by WFT and plotted in Fig. 5(a) . Since the maximum spectral derivative of the differential phase, i.e. dΔφ(λ)/dλ, always occurs within the spectral scan as indicated in Fig. 5(b), the dynamic range of our white-light spectral interferometer is only limited by the spectral bandwidth of the light source. In the present case, it is estimated to be 1.33 to 1.35, i.e. at least 10−2 RIU. Such a detection range can never be achievable with laser sources with fixed incident angle because of the highly non-linear nature of the SPR phase response.
Fig. 5 (a) (Color online) shows the differential phases of A: distilled water, B: 2% NaCl, C: 4% NaCl, D: 6% NaCl, E: 8% NaCl and F: 10% NaCl solutions; (b) is the corresponding derivatives of A to F with respect to wavelength which shows the maximum differential phase occurs at different wavelengths as A’ to F’.
To examine the LOD of our system, the wavelength specific nonlinear phase change of selected wavelengths, i.e. G: 645.15nm, H: 648.30nm, I: 653.75nm, J: 659.05nm, K: 663.60nm, L: 667.95nm, from the same experiment were plotted against the NaCl concentrations as in Fig. 6(a) . The differential phase stability of our system was experimentally verified by monitoring the refractive index fluctuation of distilled water in an hour. The standard deviation of our phase measurement was 2.5 × 10−4 radians which is approximately equal to ± 0.01°phase detection fluctuation. The LOD in the linear region is calculated via Eq. (3). It is observed in Fig. 6(a) that each wavelength has its best linear phase response at specific NaCl concentrations. For example, with 0.5% to 1% NaCl variation, the largest phase variation (0.7644 radians) occurs at the wavelength of 645.15nm (G). Our LOD is thus calculated as 2.6 × 10−7 RIU. As the NaCl concentration increases from 1% to 2%, the largest phase variation (0.8362 radians) now occurs at the wavelength of 648.35nm (H). The variation of LOD with NaCl concentration is then obtained over the range of 0% to 10% in a similar way, and is shown in Fig. 6(b). As all the values of LOD are of the order of 10−7 RIU in Fig. 6(b) for 0% to 10% NaCl concentration, which correspond to a dynamic range of 10−2 RIU, we have demonstrated that our system has attained the expected performance of high phase sensitivity with wide dynamic range. In comparison to theoretical calculation of Fig. 2, our experimental LOD is not yet optimized because the experimental incident angle is about 73° instead of 75.3° due to alignment difficulty within the Michelson interferometer.
Fig. 6 (a) (Color online) illustrates the nonlinear differential phase response of G: 645.15nm, H: 648.35nm, I: 653.75nm, J: 659.05nm, K: 663.65nm, L: 667.95nm; (b) shows the calculated Limit of Detection (LOD) of our system within the region of linear phase response of each wavelength over the 10% NaCl concentration (10−2 RIU).
5. Conclusions
We have successfully demonstrated a novel SPR sensor design based on white-light spectral interferometry by measuring the optimal differential phase with wide dynamic response. The proposed technique offers better sensitivity due to its robustness against common noise and double-pass advantage. The experimental phase detection sensitivity is already one of the best reported in literature, while it can still be improved further, e.g. by optimizing the angle of incidence as predicted by numerical simulation and replacing the LED lamp with a white-light emitting supercontinuum laser source. With the use of silver-gold sensing layer and a high-power Gaussian-beam supercontinuum source, the detected spectral phase signal should improve substantially. It is quite possible to further enhance the resolution while maintaining its attribute of wide detection range which is beneficial for biosensing applications [8], e.g. in vitro detection of cytochrome-c as a cancer biomarker that is currently under investigation.
Acknowledgement
This work was supported by a grant from City University of Hong Kong (Project No. 7002609). We also wish to acknowledge financial support of this work through the General Research Fund #411208 from the Research Grants Council of Hong Kong and CUHK-Direct Grant # 2050459.
References and links
1. J. Homola, Surface Plasmon Resonance Based Sensors, Springer Series on Chemical Sensors and Biosensors (Springer-Verlag, 2006).
2. M. Piliarik and J. Homola, “Surface plasmon resonance (SPR) sensors: approaching their limits?” Opt. Express 17(19), 16505–16517 (2009). [CrossRef] [PubMed]
3. S. Y. Wu, H. P. Ho, W. C. Law, C. Lin, and S. K. Kong, “Highly sensitive differential phase-sensitive surface plasmon resonance biosensor based on the Mach-Zehnder configuration,” Opt. Lett. 29(20), 2378–2380 (2004). [CrossRef] [PubMed]
4. A. V. Kabashin and P. I. Nikitin, “Surface plasmon resonance interferometer for bio- and chemical-sensors,” Opt. Commun. 150(1-6), 5–8 (1998). [CrossRef]
5. A. V. Kabashin, S. Patskovsky, and A. N. Grigorenko, “Phase and amplitude sensitivities in surface plasmon resonance bio and chemical sensing,” Opt. Express 17(23), 21191–21204 (2009). [CrossRef] [PubMed]
6. W. C. Law, P. Markowicz, K. T. Yong, I. Roy, A. Baev, S. Patskovsky, A. V. Kabashin, H. P. Ho, and P. N. Prasad, “Wide dynamic range phase-sensitive surface plasmon resonance biosensor based on measuring the modulation harmonics,” Biosens. Bioelectron. 23(5), 627–632 (2007). [CrossRef] [PubMed]
7. P. P. Markowicz, W. C. Law, A. Baev, P. N. Prasad, S. Patskovsky, and A. V. Kabashin, “Phase-sensitive time-modulated surface plasmon resonance polarimetry for wide dynamic range biosensing,” Opt. Express 15(4), 1745–1754 (2007). [CrossRef] [PubMed]
8. S. P. Ng, C. M. L. Wu, S. Y. Wu, H. P. Ho, and S. K. Kong, “Differential spectral phase interferometry for wide dynamic range surface plasmon resonance biosensing,” Biosens. Bioelectron. 26(4), 1593–1598 (2010). [CrossRef] [PubMed]
9. S. P. Ng, S. Y. Wu, H. P. Ho, and C. M. L. Wu, “A white-light interferometric surface plasmon resonance sensor with wide dynamic range and phase-sensitive response,” IEEE International Conference on Electron Devices and Solid-State Circuits, December 2008, HKSAR.
10. G. G. Nenninger, M. Piliarik, and J. Homola, “Data analysis for optical sensors based on spectroscopy of surface plasmons,” Meas. Sci. Technol. 13(12), 2038–2046 (2002). [CrossRef]
11. L. Mandel, and E. Wolf, Optical Coherence and Quantum Optics, (Cambridge University Press, 1995).
12. P. Hlubina, D. Ciprian, and J. Lunacek, “Spectral interferometric technique to measure the ellipsometric phase of a thin-film structure,” Opt. Lett. 34(17), 2661–2663 (2009). [CrossRef] [PubMed]
13. http://refractiveindex.info/
14. D. R. Lide ed., CRC Handbook of Chemistry and Physics, 90th ed. (CRC Press, 2010).
15. W. Yuan, H. P. Ho, C. L. Wong, S. K. Kong, and C. Lin, “Surface plasmon resonance biosensor incorporated in a Michelson interferometer with enhanced sensitivity,” IEEE Sens. J. 7(1), 70–73 (2007). [CrossRef]
16. H. P. Ho, S. Y. Wu, M. Yang, and A. C. Cheung, “Application of white light-emitting diode to surface plasmon resonance sensors,” Sens. Actuators B Chem. 80(2), 89–94 (2001). [CrossRef]
17. P. Hlubina, J. Lunacek, D. Ciprian, and R. Chlebus, “Windowed Fourier transform applied in the wavelength domain to process the spectral interference signals,” Opt. Commun. 281(9), 2349–2354 (2008). [CrossRef]
