Sensors and Actuators B 143 (2009) 381–386 Contents lists available at ScienceDirect Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb Measurements of refractive index change due to positive ions using a surface plasmon resonance sensor Hyungduk Ko ∗ , Jun Kameoka, Chin B. Su Department of Electrical and Computer Engineering, Texas A&M University, College Station, TX 77843, United States a r t i c l e i n f o Article history: Received 5 July 2009 Received in revised form 30 August 2009 Accepted 7 September 2009 Available online 2 October 2009 a b s t r a c t A 44-pass fiber optic surface plasmon resonance (SPR) sensor coupled with a field-assist capability for measurement of refractive index change due to positive and negative ions is shown. The field-assist feature forces ions to the SPR surface, causing the SPR signal response to change which reflects a decrease or increase in refractive index depending on whether positive or negative ions are being attracted to the surface. This technique offers the potential for the sensitive detection of cations and anions in a solution. Keywords: Surface plasmon resonance Multi-pass SPR sensor Fiber optic SPR sensor 1. Introduction Surface plasmon resonance (SPR) sensor has been widely studied for chemical and biological sensing in the last two decades because it allows real-time detection of biological species without labeling procedures [1–4]. It measures changes of refractive index near the surface of a thin metal layer using surface plasmons (SPs) [5]. Fundamentally, surface plasmons (SPs) are charge density waves that propagate along the surface of metals when incident p-polarized optical wave undergoes total internal reflection inside the metal coated dielectric [6]. SPR is generally implemented using the Kretschmann’s configuration [7]. At a resonance angle, the magnitude of the evanescent optical field associated with surface plasmon has a maximum value at the metal–dielectric interface and the reflectivity becomes a minimum value [6,8]. A SPR sensor utilizes this evanescent optical wave localized just above a very thin metal surface to detect molecules. Recently, we reported the 44-pass fiber optic SPR sensor that passes the detection spot 44 times, thus enhancing sensitivity by a factor of 44 [9]. Also, we reported the field-assist (applied voltage) 4-pass fiber optic sensor to detect the negatively charged particles or ions [10]. In this article, we incorporate the 44-pass feature with the field-assist method. This technique allows the detection of the refractive index change due to ions in a solution by attracting them to the SPR surface. The electric field causes the charged ions to accumulate at the Au surface, leading to an effective increase in the concentration of the ∗ Corresponding author. E-mail addresses: koh94@gmail.com (H. Ko), su@ece.tamu.edu (C.B. Su). 0925-4005/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2009.09.016 © 2009 Elsevier B.V. All rights reserved. ions at the SPR surface. We found for the first time that positive ions contribute to a decrease in refractive index while negative ions contribute to an increase in refractive index compare with the equilibrium solution. Because ions such as Na+ , K+ , H+ are important in biological processes, the detection of these ions by SPR may be a new technique for the studies of such processes. 2. Experiments Fig. 1 shows the SPR system with the fiber recirculating loop with the addition of a field-assist method. The principle and properties of the SPR system with the fiber recirculating loop to increase the number of pass, and thus the detection sensitivity, were reported in Ref. [9]. In brief, pulse generator 1 drives a laser diode (LD) to create an optical pulse train with about 5% duty cycle and 0.13 ␮s pulse width. The incident pulse is split into two pulses by fiber coupler 1 (FC1). One pulse propagates toward the SPR device by exiting port 2 of the optical circulator and the other is detected by the photodetector (PD) as shown. There are two collimators in the SPR unit. The optical beam exiting port 2 of the circulator is fed to the lower collimator and propagates toward the corner cube on the others side after reflecting off the gold surface. All fiber and collimators are single mode as multimode fiber will support many modes each exiting the fiber at different angles. Moreover, multimode fiber is incompatible with erbium-doped fiber amplifier technology. Then, the retro-reflected beam from the corner cube hits the gold surface the second time and returns to the upper collimator. Therefore, the SPR system is a two-pass configuration with the lower collimator functioning as an emitter and the upper collimator functioning as a receiver (or vice versa). 382 H. Ko et al. / Sensors and Actuators B 143 (2009) 381–386 Fig. 1. (a) An optical multi-pass SPR sensor system coupled with the field-assist method. Top: SPR, bottom: fiber loop, LD: laser diode, PD: photodiode, PC: polarization controller, and FC: fiber coupler. (b) Side view of the electrode configuration. The returned optical pulse from the SPR device is reduced in amplitude due to SPR resonance effect and back-coupling loss, therefore, the reduced amplitude needs to be amplified by the erbium-doped fiber amplifier following the electro-optic modulator (EOM). The pulse eventually returns to FC1 after one round-trip and the process repeats. Therefore, the detector (PD) detects periodic pulses that are displayed on a digital oscilloscope. The first pulse does not pass the SPR device and is irrelevant, the 2nd pulse passes the SPR twice and the third pulse pass four times, etc. The period of the periodic pulses is determined by one round-trip time around the fiber loop. The length of the delay line is not important (about 30 m in this case) as long as it guarantees that the recirculating pulses are well separated in time. Pulse generator 2 is gated and synchronized by pulse generator 1 to produce a synchronized pulse train with a longer pulse width T as shown. This gated pulse applied on the RF port of the electro-optic modulator is used for controlling its transmittance. During the time interval T when the gated pulse is applied on the electro-optic modulator, its transmittance is high and the fiber loop is closed, otherwise the loop is opened with high loss preventing pulse recirculation. This leads to the periodic opening of the EOM that prevents the fiber loop from becoming a ring laser as the optical amplifier has gain, but at the same time limits the total number of passes (pulse recirculation). Here, the EOM acts as a loss-modulating optical switch with the switch closed (low loss) when the gated electrical pulse is applied to the RF port of the EOM, otherwise the switch is opened (high loss). An important issue relates to the detrimental affect of interference fringes on the sensitivity of the multi-pass type of SPR of which the multi-pass beam overlapped at the SPR surface. If the fringe spacing (or the spot size) is smaller than the propagation decay length of the surface plasmon, one can expect a substantial decrease in SPR sensitivity in contrast with the case without interference effects. The propagation decay length of the plasmon wave along the metal surface is given by 1/(2Im(ˇ)), where ˇ, the surface plasmon wave-vector, is given by [6] k0 = 2/, and εm and εs are the gold and analyte dielectric constant, respectively. Using εm = (0.2 + i10.2)2 for gold and εs = 1.31592 for water, the propagation decay length is about 250 ␮m for a wavelength, , of 1.53 ␮m used in this experiment. The plasmon wave cannot freely propagate because the interference fringe spacing is much less than 250 ␮m. Therefore, in this work, a corner cube was used for retro-reflection instead of a mirror, assuring that the multipass beam does not overlap. In addition, as mentioned previously, two collimators side-by-side are used for delivering and receiving light that passes the SPR surface instead of one collimator. The corner cube retro-reflect the light from one collimator back into the other collimator. With this method the light spot at the SPR surface does not overlap, avoiding the possible interference effects that occur with the use of one collimator and a mirror reflector scheme as reported in Ref. [9]. However, if the light source has low coherence then the one collimator and one reflecting mirror scheme may still be used if the optical path length spacing between the mirror and the light spot at the SPR surface is longer than the coherence length of the light source. To apply an electric field to the analyte solution, another Au-coated glass wafer is used as the top electrode. The bottom electrode is the SPR surface. Poly(dimethylsiloxane) (PDMS) sheets with 1 mm thickness are fabricated with two holes that function as solution wells. The PDMS sheet is placed between the bottom and top electrodes. Analyte solution is dropped into the two holes on the PDMS sheets, and the solution is covered by the top electrode. The bottom electrode is electrically grounded. Therefore, when a positive voltage is applied to the top electrode, positive ions are collected onto the bottom SPR surface. The light beam hits the region of the solution well as shown in Fig. 1(b). In this paper a light source wavelength of 1.53 ␮m compatible with erbium-doped fiber amplifier technology is used. Compared to the commonly used wavelengths of 0.67 and 0.85 ␮m, the resonant profile is sharper and the reflectivity dip is shallower at 1.53 ␮m wavelength, as verified in Ref. [11]. 3. Results and discussion ˇ = k0  εm εs εm + εs (1) The one-pass SPR reflectivity profiles versus incident angle for DI water as reference solution is shown in Fig. 2. The presence of H. Ko et al. / Sensors and Actuators B 143 (2009) 381–386 383 Fig. 2. SPR reflectivity curve as a function of angle. The bias angle is set below the resonant angle as depicted. a substance with larger refractive index than the water’s refractive index shifts the resonance angle towards larger values. Thus, by setting the bias angle below the resonance angle, the reflected optical power increases with the solution’s index. To investigate and calibrate the 44-pass response with respect to changes of refractive index, various concentrations of salt solutions in DI water are prepared and measured. The corresponding pulse signal amplitude of the 22nd pulse associated with the various salt concentrations is measured and the result plotted in Fig. 3. Since each pulse corresponds to two passes and the total number of pulses is 22, the total number of passes is 44. Fig. 3 shows that the differential change in amplitude increases with the number of passes and salt concentration. The solution’s index increases with the increase of salt concentration, which causes the reflected optical power to increase. Since the index change can be calculated according to the salt concentration, we can calibrate the 44-pass response with respect to index change using salt solution. Fig. 3 shows that there is a gradual increase in the baseline signal with time, which is due to amplified spontaneous emission associated with the erbium-doped fiber amplifier itself. If more pass is allowed, lasing action eventually sets in, i.e., if the EOM gate is left closed for too long, the setup eventually becomes a ring laser, destroying the SPR function. Thus, the ultimate number of pass is limited by lasing effects. The signal-to noise ratio degrades as the number of pass increases due to increased spontaneous emission. The noise in erbium-doped fiber system is dominated by the so-called signalspontaneous beat noise and the spontaneous-spontaneous beat noise as described in Eqs. (7.42) and (7.43) in Ref. [13]. The total mean square current fluctuation (ıI)2 is approximately given by the sum of the signal–spontaneous beat noise and the Fig. 4. (a) Normalized pulse amplitude for 44 passes versus salt concentration. (b) Normalized pulse amplitude for 44 passes versus refractive index change corresponding to the salt concentration. spontaneous–spontaneous beat noise [13]: 2 (ıI) = IS Isp (2) where IS and Isp describe the average detected signal and the spontaneous emission current, respectively. Be describes the detection electronics bandwidth (108 Hz) and B0 describes the optical filter bandwidth (1 nm, 1.28 × 1011 Hz) placed within the fiber loop (Fig. 1). The optical filter in the loop is very important for reducing Isp . The shot noise contribution due to IS and Isp and the thermal noise due to the detection electronics are neglected because they are appreciably smaller in magnitude. The signal-to-noise ratio S/N is given by, S = N Fig. 3. Pulse amplitude versus number of passes for DI water, 340 ␮M salt, 680 ␮M and 1.02 mM salt. Note that each pulse involves passing the SPR device twice. Therefore, 22 pulses correspond to 44 passes. Be 2 Be + Isp B0 B0 IS  (ıI) 2 (3) For the 44 passes describes in Fig. 3, the spontaneous emission current Isp (measured voltage divided by the scope’s input resistance of 2.5 × 106 ), described by the base line, increases by a factor of two from 2-pass to 44-pass. Thus, for example, the S/N degrades by no more than a factor of two from 2-pass to 44-pass for the 1.02 mM case shown in Fig. 3. Specifically, with numbers given above, the S/N for 2-pass is 108 and 69 for 44-pass. If the gain of the erbium-doped fiber amplifier is adjusted too high, Isp can increase dramatically eventually leading to lasing of the ring at the optical filter frequency and dramatically degrading the signal S/N. However, if the gain is too low, the signal decreases as the number of pass increases. Thus, one has to adjust the gain just right. From Fig. 3, the normalized 44-pass pulse signal with respect to salt concentration and its corresponding index’s change are plotted in Fig. 4. As shown in Fig. 4(a), the normalized pulse signal increase 384 H. Ko et al. / Sensors and Actuators B 143 (2009) 381–386 field in the optical regime. According to Eq. (3.71) of Ref. [14], the response of the bounded electron to the optical field give rise to a refractive index n that can approximately be described by, n2 = 1 + Fig. 5. Time-dependent SPR signals for DI water. (a) 2-Pass (2nd pulse) and (b) 44-pass (22nd pulse). The applied voltage is 0.8 V. is linearly proportional to salt concentration. The normalized pulse amplitude corresponding to the index change is shown in Fig. 4(b). The conversion from salt concentration to refractive index change, n, is by the formula: nw + n = ns x + (1 − x)nw [12], where ns and nw are the refractive index of salt and water, respectively, and x is the salt weight fraction. The refractive index change per the normalized pulse amplitude change, 1.13 × 10−5 gives the best fit to the measured data in Fig. 4(b). This allows us to predict the unknown index change of a solution. Previously, we reported a time-dependent SPR signal response for the attraction of negatively charged ions to the SPR surface. It has been demonstrated that the attraction of the negatively charged ions to the SPR surface lead to an increase in the SPR signal if the bias angle is below the resonance angle as indicated in Fig. 2. The SPR signal increase due to the applied field indicates an increase of the refractive index due to the negative ions. In this article, a time-dependent SPR signal response due to the attraction of cations (positive ions) to the SPR surface is investigated. It will be shown that the attraction of positive ions to the SPR surface causes the signal to decrease indicating a decrease in the refractive index in contrast with an increase in index for negative ions. Furthermore, the response due to positive ions is much smaller than for negative ions, therefore, requiring more SPR passes than Ref. [10] to enhance sensitivity. First, the SPR signal response to a 0.8 V step voltage for DI water is measured as shown in Fig. 5 for both 2 passes and 44 passes. The 2pass data is taken from the amplitude change of the 2nd pulse from among the 22 pulses (note that the first do not pass through the SPR so is not considered). The 44-pass data is taken from the 22nd pulse. The decrease in signal is presumably due to the attraction of the SPR surfaces of H+ ions, which at a pH of 7 has a concentration of 10−7 M. The signal suddenly increases abruptly at about 178 s because the voltage is suddenly switched to the opposite polarity to attract the negative OH− ions to the SPR surface. The attraction of negative ion is not of interest in this paper as this has been reported in Ref. [10]. Here, the SPR signal decrease means that the refractive index of the solution is locally reduced in the detection region. To our knowledge, the refractive index associated with negative and positive atomic ions have never been reported, thus, here we can only make some qualitative conjecture as to why negative ions increase while positive ions decrease the refractive index with respect to the neutral environment, and that the magnitude change is much larger for negative ions compared with positive ions. The rational is as follows: the magnitude of the refractive index of materials in the optical regime is described by the response of the bounded or the free electron to an exciting optical field as described in Chapters 3 and 4 of Ref. [14]. This is because only electrons and not the atomic ions have small enough mass to react to the fast oscillating electric Nq2 f ε0 me ω2 − ω2 0 (4) Important parameters we need to make use of are the optical frequency ω, the electron density N and the resonance frequency ω0 of the particular bounded electron system. More tightly bounded electrons give higher ω0 , which, for most materials are in the ultraviolet frequencies (Table 4.3 of Ref. [14]). The attraction of positive ions to the detection surface merely decrease N and therefore decrease n, while the attraction of negative ions increase N and may also simultaneously decrease ω0 because the outer electron in the negative ion (Cl− in NaCl, for example) is not as tightly bounded compared with the inner electrons. Since ω0 ≫ ω, a decrease of ω0 can potentially dramatically increase n, according to Eq. (3). Thus, the detection of negative ions increases the refractive index while the detection of positive ions decreases the refractive index with respect to the neutral environment in the absence of an applied voltage. It should be noted that SPR signal change induced by the applied voltage is a function of the refractive index change, and is a function of the concentration of cations (positive ions) in aqueous solution. Therefore, the refractive index of a concentration of cations in aqueous solution can be estimated because the refractive index change per the fractional signal change can be predicted from Fig. 4(b). Moreover, 44 passes offer much better sensitivity than 2 passes, demonstrating that our device has good sensitivity to measure the index’s change due to the attraction of cations. The estimated refractive index decrease using the calibration data of Fig. 4a and b indicates that the attraction of H+ cause the refractive index to decrease by roughly 2 × 10−6 at an applied voltage of 0.8 V. 1 M KOH and 1 M NaOH are then used as test solutions for measuring the refractive index of cations in aqueous solution to see whether the attraction of K+ and Na+ also cause the refractive index to decrease. These solutions are chosen because K+ and Na+ ions are important to biological processes. The phenomenon reported here may provide a new method for studying ion exchange across cell membrane in cells. A time-dependent 44-pass SPR signal response for the solution are measured and compared to DI water. For comparison, the one-pass reflectivity versus the incident angle is verified for the KOH and NaOH solutions, respectively, then the bias point is set to the same voltage level as that of DI water. Since KOH and NaOH is a strong base in aqueous solution and their degree of electrolytic dissociation are almost equal, we assume that 1 M of K+ and Na+ ions exist in each aqueous solution. As predicted, Fig. 6 indicates that the SPR signal is decreased for both KOH and NaOH Fig. 6. Time-dependence of the amplitude of the 22nd pulse (44 pass) for DI water, 1 M KOH and 1 M NaOH. The applied voltage is 0.8 V. H. Ko et al. / Sensors and Actuators B 143 (2009) 381–386 385 the applied field, the current through the solution well filled with DI water was measured as shown in Fig. 7(a). The current was determined by measuring the voltage across a 20 k resistor that was connected in series with the solution wells. From Fig. 7, we can calculate the energy generated from the joule heating using  E = V I dt where V is the applied step voltage of 0.8 V. The calculated energy E is 6.1296 × 10−5 for 120 s. The specific heat capacity of DI water, Csp,water is 4.186 J g−1 ◦ C−1 and the amount of water in the solution well was 0.24 g. As a result, the increase in temperature due to the applied field is about 6.1 × 10−5 ◦ C for 120 s since the energy needed to increase the temperature by 1 ◦ C is 1.00464 J. This results in the change in refractive index of about −6 × 10−9 , which is not significant with the local index change of about −10−6 at the SPR surface due to the attraction of positive ions. In the same way, the refractive index change in 1 M NaOH solution due to the joule heating was calculated from Fig. 7(b) and it was about −2.0 × 10−7 less than the local index change due to the attraction of Na+ ions to the SPR surface (specific heat capacity, Csp,1 M NaOH is 3.975 J g−1 K−1 ). Therefore, joule heating is not a contributing factor to the results shown in Figs. 6 and 7. 4. Conclusion Fig. 7. Current flowing in (a) DI water and (b) 1 M NaOH upon the turn-on of the step voltage of 0.8 V. upon applying a positive voltage to the top electrode. This is due to the attraction to the SPR surface of K+ and Na+ ions. The degree of signal drop in response to the external field is proportional to the refractive index drop of cation in aqueous solution. Moreover, according to Fig. 4(b), the estimated refractive index decrease is of the order of about a few times 10−6 . In Fig. 6, the fractional pulse amplitude changes are about 0.15, 0.27 and 0.45 for DI water, 1 M KOH solution and 1 M NaOH solution, respectively. However, these values may vary by as much as 30%, if the experiment is repeated by recycling the applied voltage. The variation is probably due to residual ions or molecules sticking to the SPR surface so that each run is not exactly the same. But the downward signal trend upon the application of a positive voltage shown in Fig. 6 is always the same for repeated experiments. Moreover, if more accurate measurements were to be attempted, some reference calibration methods or temperature controlled must also be implemented as the refractive index of water change by about 10−4 per degree C. Here, we only attempt to demonstrate that the attraction of positive ions to the SPR surface cause the refractive index to decrease. We choose to apply a voltage of 0.8 V so that it is below the water electrolysis voltage of 1.23 V. Also a lower voltage protects the integrity of the gold surface for further use. If a higher voltage is applied, the signal magnitude will be bigger. Additionally, the pulse signals for the solutions increase dramatically once the reversed voltage is applied, which is due to the attraction of OH− ions to the SPR surface. For the case of KOH and NaOH solutions, the stronger increase in the signal compared with DI water is shown. This should be due to the attraction of much higher concentration of OH− ions of 1 M of KOH or NaOH than the OH− ions of DI water. Joule heating due to the applied field can also cause the signal to decrease because the refractive index of water decreases with temperature. To investigate the temperature increase due to In conclusion, a field-assist 44-pass SPR fiber optic technique for measurement of positive ions is demonstrated. Detection of positive ions gives a SPR signal that indicates a decrease in the refractive index, while the detection of negative ions gives a signal that indicates an increase in the refractive index. The detection of positive ions by the SPR technique may be a new method for studying ion exchange in cells. The multi-pass SPR sensor system offers sensitivity enhancement by a factor depending on the number of passes. The technique offers a method that can measure the refractive index with improved sensitivity. References [1] B. Liedberg, I. Lundstrom, E. Stenberg, Principles of biosensing with an extended coupling matrix and surface plasmon resonance, Sens. Actuator B 11 (1993) 63–72. [2] X. Liu, D. Song, Q. Zhang, Y. Tian, L. Ding, H. Zhang, Wavelength-modulation surface plasmon resonance sensor, Trends Anal. Chem. 24 (2005) 887–893. [3] K. Matsubara, S. Kawata, S. Minami, Optical chemical sensor based on surface plasmon measurement, Appl. Opt. 27 (1988) 1160–1163. [4] J. Homola, S.S. Yee, G. Gauglitz, Surface plasmon resonance sensors: review, Sens. Actuator B 54 (1999) 3–15. [5] E. Fujii, T. Koike, K. Nakamura, S. Sasaki, et al., Application of an absorptionbased surface plasmon resonance principle to the development of SPR ammonium ion and enzyme sensors, Anal. Chem. 74 (2002) 6106–6110. [6] W.L. Barnes, A. Dereux, T.W. Ebbesen, Surface plasmon subwavelength optics, Nature (London) 424 (2003) 824–830. [7] E. Kretschmann, Determination of optical constants of metal by excitation of surface plasmons, Z. Phys. 241 (1971) 313–324. [8] A. Otto, Excitation of nonradiative surface plasma waves in silver by the method of frustrated total reflection, Z. Phys. 216 (1968) 398–410. [9] C.B. Su, J. Kameoka, Forty-four pass fibre-optic loop for improving the sensitivity of surface plasmon resonance sensors, Meas. Sci. Technol. 19 (2008) 015204. [10] H. Ko, C.B. Su, J. Kameoka, Chracterization of a field-assist multipass fibre optic surface plasmon resonance sensor, Meas. Sci. Technol. 18 (2007) 2938–2942. [11] C.B. Su, J. Kameoka, B. Ilic, K.H. Chu, K.A. Chang, Properties of an optical multipass surface plasmon resonance technique, Appl. Phys. Lett. 89 (2006) 071101. [12] K.A. Chang, H.J. Lim, C.B. Su, A fibre optic fresnel ratio meter for measurements of solute concentration and refractive index change in fluids, Meas. Sci. Technol. 13 (2002) 1962–1965. [13] P.C. Becker, N.A. Olsson, J.R. Simpson, Erbium-Doped Fiber Amplifiers, Fundamental and Technology, Academic Press, San Diego, 1999. [14] E. Hecht, Optics, Addison Wesley, San Francisco, 2002. Biographies Hyungduk Ko received his Ph.D. in Electrical Engineering from Texas A&M University in 2009. Currently, he is a senior researcher in Samsung LED Co. Ltd. His research interests include nanophotonics and optics, plasmonic-assistedoptoelectronic 386 H. Ko et al. / Sensors and Actuators B 143 (2009) 381–386 device, solar cells and imaging and surface plasmon resonance (SPR) fiber optical sensors. Jun Kameoka received his Ph.D. from Cornell University in 2002. Since 2004, he has been an assistant professor in Texas A&M University. His research interests include bio-nanomachining, nanostructure science and engineering, nanosensors and molecular manipulation, micro and nanofluidics and bio-nanohybrid devices for medical applications. Chin B. Su received his Ph.D. in Physics from Brandeis University in 1979. In 1978–1986 he was a member of technical staff in GTE laboratories. In 1986–1987, he was a project team leader in Rockwell International Corp. Since 1987, he has been on the faculty of the Department of Electrical Engineering at Texas A&M University, currently, he is a full professor in Texas A&M University. His research interest is surface plasmon resonance sensor for bio-detection.