1. Introduction

The principle interest that scientists have in the use of surface plasmon resonance (SPR) biosensors, which are now widely used in the fields of life science and drug discovery, is that it provides a highly sensitive platform to study biomolecular interactions in real time [1]. The feature of a SPR biosensor is the ability to monitor biomolecular interactions via the intensity variation [2] or phase shift [3] of a reflected laser beam introduced by surface plasma wave (SPW) excitation within the SPR device. This is because the SPW property is critically dependent upon the change in effective refractive index (Δn_eff), including the variations in refractive index and thickness of active medium within the active region of SPW from metal surface [4,5]. Therefore the localized distribution of the electric field of SPW in the vicinity of metal/dielectric interface is crucial to the detection sensitivity. Then, the Δn_eff can be expressed by (Δn)_eff=(∂n_eff/∂n)·Δn+(∂n_eff/∂d)·Δd where Δn and Δd are the changes in refractive index and the thickness of the dielectric medium due to binding events. The detection of Δn_eff by way of measuring the attenuated intensity or phase shift of the reflected P polarized wave (TM wave) results SPR biosensor being able to monitor biomolecular interactions in real time without the use of labeling dye molecule and this allows applications in the area of kinetic analysis such as association rate constant, dissociation rate constant and equilibrium dissociation constant for various biomolecular interaction including protein, lipid, nucleic acid and other small molecules [6,7].

Generally, SPR biosensors can be divided into two categories, those that use interferometric methods [3–5] and those that use non-interferometric methods [4,8]. In the non-interferometric method, Δn_eff is measured either by tracing the resonance angle of the SPW in order to detect the minimum optical intensity of the reflected P wave or by selecting the on-resonance wavelength, which is able to excite the SPW correctly [1,4]. In general, Δn_eff ≈10^-6 to 10^-7 RIU (refractive index unit) is the sensitivity obtained by noninterferometric methods [1,4] in which a large bandwidth of the noise power spectrum is detected. These properties are in contrast to interferometric methods where the method detects attenuation of the amplitude or the phase shift of the reflected P polarized heterodyne signal [3–5]. These methods are based on synchronized detection and have a narrow bandwidth of the noise power spectrum, which results in a high signal-to-noise ratio (SNR) and a high modulation index (MI) of the heterodyne signal. Recently, Δn_eff≅10^-8 RIU has been demonstrated experimentally when using the phase-sensitive detection method with a phase modulated optical heterodyne interferometer [3]. However, a limited dynamic range for the phase response induced by the SPW excitation is resulted. Thus, the phase-sensitive method works only for small biomolecules or low experimental concentration [3,9].

Previously, we have developed an optical heterodyne SPR biosensor [5] using a Zeeman He-Ne laser that outputs a pair of correlated and orthogonal linear polarized P and S light waves with different temporal frequencies, which excites a pair of correlated surface plasma waves simultaneously on the interface of metal/dielectric medium of a SPR device. According to Kuo’s et al. [5] analysis, this paired SPW optical heterodyne biosensor forming the SPR device is capable of improving the detection sensitivity significantly compared to a conventional SPR biosensor in which a single SPW is excited. As a result, the rate of change of the attenuated amplitude of the P polarized heterodyne signal versus Δn_eff is apparently enhanced. At the same time, a wide dynamic range for Δn_eff is obtained as well because the attenuated amplitude rather than the intensity of the P polarized heterodyne signal is measured in this arrangement. However, the linear birefringent effect of the laser cavity of the Zeeman He-Ne laser produces elliptical polarization and non-orthogonality of these two linear polarizations of the P and S waves in the laser beam [10]. This induces decorrelation between paired SPWs partially and then obviously results in a degradation of detection sensitivity because of a lowering on SNR and MI of the detected heterodyne signal. In addition, thermal disturbance within the laser cavity induces laser intensity fluctuation and laser frequency noise at the same time. Therefore, we propose a method that is able to reduce these system noises and further enhances the detection sensitivity of paired SPW biosensor.

2. Principle

In order to improve the amplitude and frequency stability of the laser beam, an amplituderatio algorithm is introduced into this paired SPW biosensor whereby the excess noise of laser intensity fluctuation can be effectively reduced [11,12]. Simultaneously, if a common-path polarized heterodyne interferometer is used, then the common phase noise including the background phase noise and the frequency noise are inherently cancelled out as well. Therefore, a pair of S polarized waves (TE wave), which are independent of SPW excitation, is required in order to normalize the P polarized signal in real time. Furthermore, a pair of S polarized waves propagating with a common-path along with the P waves is necessary too. Theoretically, there should be no amplitude attenuation and phase shifted of the S waves when they are totally reflected by SPR device [13].

 

Fig. 1. Schematic of the amplitude-sensitive PSPWB: λ/2 : a half-wave plate; AO1, AO2: acoustic-optic modulators; D1, D2: drivers; PBS: a polarized beam splitter; BS1, BS2: beam splitters; M1, M2: mirrors; He-Ne: He-Ne laser; RC: reaction chamber; D_p and Ds: photo detectors; BF1 and BF2: band pass filters; LIA: lock-in amplifier.

Download Full Size | PDF

Thus, a polarized common-path interferometer is setup for this novel paired SPW biosensor (PSPWB). It is able to reduce the excess noise from laser intensity fluctuation and the background noise at the same time, that the PSPWB can decrease the effect of environmental disturbance such as temperature variation in the reaction chamber of SPR device. This is very important to retain the amplitude stability in this novel PSPWB to ensure high detection sensitivity.

In Fig. 1, a frequency stabilized linear polarized He-Ne laser is introduced into a polarized common-path optical heterodyne interferometer in which the two pairs of highly spatial and temporal correlated linear polarized light waves, namely the P polarized ones, (P₁+P₂) and the S polarized ones, (S₁+S₂) are generated by the interferometer simultaneously. The setup is different from the previously developed paired SPW biosensor using a Zeeman laser [5] in which the reference beam is not common-path propagation to the signal beam of paired P waves in the reaction chamber of SPR device.

The half wave plate (HWP), in Fig. 1, rotates the linear polarization of the laser beam by 45° to the x axis so that an equal amplitude of P and S waves is produced. A fixed frequency shift of ω ₁ to P₁ and S₁ waves in the upper channel of the interferometer is introduced by an acoustic modulator AOM1 while e a ω ₂ to P₂ and S₂ waves in the lower channel of the interferometer is introduced by AOM2 simultaneously. Thereafter, a beam splitter (BS) is used to recombine the P₁ and S₁, and P₂ and S₂ waves together so that two pairs of correlated linear polarized light waves, P₁+P₂ and S₁+S₂, are produced in the laser beam. Both pairs of P polarized and S polarized light waves have the same frequency difference Δω=ω ₁-ω ₂ for the heterodyne signals. Then, the laser beam is made incident into a SPR device with a Kretschmann’s configuration as shown in Fig. 1 such that a pair of SPWs is excited by the paired P polarized waves on the metal/dielectric interface via attenuated total reflection (ATR). In the meantime, the paired S waves, which show a common-path propagation along with the paired P waves, are totally reflected by SPR device simultaneously. Then, a polarized beam splitter (PBS) separates the P₁+P₂ waves and S₁+S₂ waves. The separated signals are optically heterodyned at the photo detectors D_p and D_s, respectively.

The reflected P and S polarized heterodyne signals in this setup can be expressed by

The DC terms in Eqs. (1) and (2) is filtered out by use of a band pass filter (BPF) at a center frequency $\Delta \omega ·(A_{P_1}^{\prime },A_{P_2}^{\prime })$ are the attenuated amplitudes of the reflected P₁ and P₂ waves respectively, while $(A_{S_1},A_{S_2})$ are the amplitudes of S₁ and S₂ waves. Δφ_P and Δφ_S are the phase differences of the reflected P₁ and P₂ and the reflected S₁ and S₂ waves respectively. Furthermore, the ratio of the amplitudes χ of the two polarized heterodyne signals, $\chi =(A_{P_1}^{\prime }·A_{P_2}^{\prime }⁄A_{S_1}·A_{S_2})$ is measured by using a lock-in amplifier. The S wave becomes the reference signal that normalizes the attenuated P wave in order to reduce the excess noise due to laser intensity fluctuation. Simultaneously, the common background phase noise induced by temperature variation in the reaction chamber of SPR device together with laser frequency noise are cancelled inherently. Then, the conversion from phase noise into amplitude noise within the PSPWB is prevented [14].

Figure 1 shows the configuration of PSPWB, which basically is built on a polarized common-path Mach-Zehnder optical heterodyne interferometer. Two acousto-optic phase modulators (AOMs) are inserted into the interferometer, where AOM1 was driven at a frequency ω ₁=80.000MHz while AOM2 was driven at a frequency ω ₂=80.033MHz simultaneously, with a result that Δω=ω ₁-ω ₂=33kHz in this arrangement. A beam splitter (BS2) recombines the laser beams from the two channels of the interferometer and then the beam is incident into a SPR device at an angle θ close to the resonance angle θ_R in order to excite the SPW.

Because the P₁ and P₂ waves correlate with each other in the temporal and spatial domain and their frequency difference is 33kHz, then the amplitude attenuation and phase shift of the two reflected P waves induced by exciting a pair of SPWs have approximately equal response. Thus, the phase difference of paired P waves induced by SPW excitation is Δφ_spw≅0. Meanwhile, the phase difference Δφ_l produced by optical path length l of the P₁ and P₂ waves is negligible, i.e. Δφ_l≅(Δω/c)nl≅0, where n is the refractive index, and c is the speed of light. As a result, the total phase difference of the P heterodyne signal in Eq. (1)becomes Δφ_p=Δφ_spw+Δφ_l≈0. Similarly, Δφ_S≅0 is also concluded because S wave is independent of SPW excitation.

3. Experimental results

To characterize the detection sensitivity of this PSPWB, a series of sucrose (Sigma, St. Louis, MO) solutions of different weight percent (%) concentrations from 0.001% to 10% were tested. The weight percent concentration is expressed as (weight of solute weight of solution)×100. Another commonly used percentage concentration is weight volume percent concentration which is expressed as (weight of solute volume of solution)×100. Thus, 10% by weight of sucrose solution equals to 111mg/ml of weight volume concentration [15]. The sucrose solution flowed continuously over the sensor surface of the sensor chip (Au-biochip, Biacore AB, Uppsala, Sweden) during testing. The sensor chip has a bare gold surface of 45nm in thickness, which is designed to avoid any influence on refractive index caused by a linker layer. Fig. 2 shows the experimental results of the measurement on χ, the amplitude ratio of P and S heterodyne signals, using a lock-in amplifier. In Fig. 2(a), the concentration of sucrose solution ranges from 1.0% to 10%; and in Fig. 2(b) the range from 0.01% to 1.0% under the same conditions and in Fig. 2(c) the range from 0.001% to 0.005% accordingly. The corresponding index of refraction of each sucrose concentration can be precisely calculated referring to a refractive index table of sucrose solution [16].

 

Fig. 2. Detected amplitude obtained from the PSPWB at different wt% of sucrose-water solution. (a) concentrations in a range of 1–10%, (b) concentrations in a range of 0.1–1.0%, (c) concentration in a range of 0.001–0.005 %.

Download Full Size | PDF

In Fig. 2, the time response of different concentrations of sucrose solutions ranging between 0.001% and 10% demonstrates a dynamic range of 104. The wide dynamic range of PSPWB is because of the amplitude detection of heterodyne signal. According to the refractive index table of sucrose solution, the concentration of 0.001% of sucrose solution is equivalent to the refractive index change of Δn=1.4×10^-6 RIU (refractive index unit) relative to pure water. Since there is no thickness variation involved in this experiment that Δn_eff=Δn is satisfied. The detection sensitivity of PSPWB can be calculated from the relationship of δ(Δn)=Δn/SNR where SNR stands for signal to noise ratio of heterodyne signal. The SNR of 0.001% sucrose solution is measured by the ratio of signal level to noise level as drawn in Fig. 2(c) in which V_N stands for the magnitude of noise level, V_S stands for the magnitude of signal level of 0.001% sucrose solution. In this way SNR≅4 is acquired, and that the detection sensitivity of PSPWB becomes δ(Δn)=3.5×10^-7 RIU experimentally. If detecting protein-protein interaction of mouse IgG/anti-mouse IgG, the detection sensitivity of PSPWB is calculated as δ(Δn_eff)≅8×10^-10 RIU based on the molecular weight ratio of IgG (150kDa) and sucrose (342Da).

 

Fig. 3. The time response of glycerin-water solutions with various percentages by weight concentration. Glycerin-water concentration in the range of (a)1.0–10% and (b) 0.01–0.5%

Download Full Size | PDF

The sensitivity of the PSPWB was also tested using a glycerin-water solution over concentrations from 0.01% to 10%, as shown in Fig. 3. Under these circumstances, the detection sensitivity of PSPWB is δ(Δn)=3.0×10^-6 RIU at 0.01% (SNR≈4 in Fig. 3(b)). The difference in the detection sensitivity between sucrose solution and glycerin solution experiments is due to the variation in molecular weight of these two solutes. If the detection sensitivity derived from 0.01% glycerin-water solution is converted into sucrose solution detection, the detection sensitivity of PSPWB is δ(Δn)=7.6×10^-7 RIU, which is consistent with that the result from 0.001% of sucrose-water solution. In the same way, for detecting protein-protein interaction of mouse IgG/anti-mouse IgG, the detection sensitivity of PSPWB is calculated as δ(Δn_eff)≅2×10^-9 RIU based on the ratio of the molecular weight of IgG (150kDa) and glycerin (92Da). This result agrees with the sensitivity of detection via sucrose solution in previous discussion.

In comparison, Wu et al. [3] experimentally demonstrated the sensitivity of δ(Δn_eff)≅5.5×10^-8 RIU by testing glycerin-water solution of different concentrations, 0.1% to 8%, by using a phase modulation polarized heterodyne interferometer via phase-sensitive detection method. However, a 0.01° phase stability of the interferometer is required. In the meantime, when the same system was used for bovine serum albumin (BSA) detection, the lowest measurable concentration became 7.4ng/ml under the same condition of 0.01° resolution of phase measurement. This is because of the limited dynamic range of the phasesensitive method of which the detection sensitivity drops so rapidly because of the saturation of the phase response in the SPR biosensor.

In order to directly verify the detection sensitivity of the PSPWB for biomolecular interaction, an experiment using IgG/anti-mouse IgG interaction was carried out. First, the sensor surface on CM5 sensor chip (Biacore AB, Sweden) was synthesized with rabbit antimouse IgG (40µg/ml) according to the amine coupling method [17]. The carboxyl groups of sensor chip was activated to form reactive N-hydroxysuccinimide esters using a solution of 0.2M N-ethyl-N’-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC) and 0.05M Nhydroxysuccinimide (NHS) in water. After activation, rabbit anti-mouse IgG in 10mM sodium acetate, pH 5.0, was immobilized via reaction of its nucleophilic groups. Excess esters were then deactivated using 1M ethanolamine hydrochloride adjusted to pH 8.5 with sodium hydroxide, which also desalted loosely bound protein. Then a covalent attachment of antimouse IgG onto the sensor surface was performed following the activation, coupling and deactivation steps. Figure 4 shows the sensogram for immobilization of anti-mouse IgG at the concentration of 40µg/ml on CM5 sensor chip.

 

Fig. 4. Sensogram of the immobilization of rabbit anti-mouse IgG by use of the amine coupling method on CM5 sensor chip.

Download Full Size | PDF

After the coupling procedure, the binding process between 10pg/ml (or 67fM) mouse IgG and the immobilized anti-mouse IgG was monitored. Figure 5 shows the response curve in real time. A high SNR (=24) is clearly seen in Fig. 5.

 

Fig. 5. Time response of 10pg/ml mouse IgG interacting with immobilized rabbit anti-mouse IgG on CM5 biochip

Download Full Size | PDF

The sensitivity of PSPWB can be also compared with a conventional SPR biosensor of tracing the displacement of the resonance angle which shows a minimum detectable concentration at 15ng/ml of BSA interacting with carboxymethylated dextran layer on the gold film of CM5 biochip [18]. Thus, as shown in Fig. 5, the detection sensitivity of PSPWB is 7 × 10² times higher than that of a conventional SPR biosensor because of the ratio on of molecular weight of IgG (150kDa) and BSA (68kDa). This implies that the performance of PSPWB is equivalent to the shot-noise limited detection of a conventional SPR sensor at a minimum concentration of 15 pg/ml of BSA based on shot-noise-limited calculation [18].

In addition, a direct comparison was made between PSPWB and Young’s interferometer of a phase-sensitive biosensor in which the number of fringe shifted is measured during the detection [19]. In the experiment, a 50ng/ml of IgG interacting with immobilized protein-G was measured by PSPWB. As described previously, protein-G of 50µg/ml in 10mM sodium acetate, pH 4.0, was immobilized onto the sensor surface of CM5 by amine coupling method [17]. Then, a 50ng/ml mouse IgG interacting with the immobilized protein-G was monitored in real time. The result is shown in Fig. 6 where (SNR)_PSPWB=120 is obtained by PSPWB versus (SNR)_Young’s=3 obtained by Young’s interferometer method under the same condition [8]. Thus, an improvement factor α=(SNR)_PSPWB/(SNR)_Young’s≅40 is calculated in favor of PSPWB on detection sensitivity. This implies that the PSPWB is equivalent to δ(Δn_eff)≅2.3×10^-9 RIU in mouse-IgG/protein-G interaction compared to given sensitivity of δ(Δn_eff)≅9×10^-8 RIU by Young’s interferometric biosensor at the same conditions.

 

Fig. 6. A 50ng/ml mouse IgG interacting with immobilized protein-G in PBS buffer solution. (SNR=120)

Download Full Size | PDF

Meanwhile, according to Kuo’s et al. experimental result [5], a large dynamic range from 25 to 1000ng/ml of IgG/anti-IgG interaction has been achieved using PSPWB previously. In this study, we have further improved that PSPWB can detect much lower concentration of IgG down to 10pg/ml successfully. Therefore a dynamic range of 10⁵ by this improved PSPWB is achieved.

4. Conclusions and discussion

In order to retain the high sensitivity performance of the PSPWB when detecting effective refractive index variation, following conditions are required in this novel PSPWB. They are (1) two pairs of highly correlated, and orthogonal linear polarizations, the P₁+P₂ waves and S₁+S₂ waves, at same beat frequency; (2) the amplitude-ratio algorithm for the P and S polarized heterodyne signals in which P polarization is the signal and S polarization is the reference beam; (3) a common-path propagation of P and S polarized heterodyne signals in the interferometer; (4) a highly spatial and temporal correlation of a pair of P waves and a pair of S waves in the interferometer; (5) a high surface quality of gold thin film in the SPR device. All these requirements result the PSPWB able to excite a pair of highly correlated SPWs simultaneously thus, gives rise to a high SNR and MI of the heterodyne signals. At the same time, a reduction in the common background noise in the amplitude and phase of P and S heterodyne signals occurs. Thus, the measurement of the variations of normalized amplitude χ of P polarized heterodyne signal versus effective refractive index n_eff, i.e., (Δχ/Δn_eff), becomes much more sensitive than that of single SPW excitation in a conventional SPR biosensor [18]. Moreover, to integrate the property of the paired S waves propagating along a common-path with the paired P waves in PSPWB with the amplitude-ratio algorithm results that the PSPWB is able to perform on detection sensitivity equivalent to the shot-noise-limited of a conventional SPR biosensor. This implies PSPWB that is capable of producing detection sensitivity at δ(Δn_eff)≅10^-9 RIU when studying protein-protein interactions.

It is clear that high spatial and temporal correlations between a pair of SPWs excited on the gold surface in SPR device are required in order to produce the high sensitivity of PSPWB. Theoretically, the beam deviation angle between two P waves or two S waves within several tenth of micro-degree is required such that a high heterodyne efficiency of the signal is obtained [20]. Or, a degradation of the sensitivity of PSPWB is resulted. To our knowledge, this novel PSPWB presents the highest detection sensitivity on Δn_eff compared to conventional SPR biosensors of interferometric and non-interferometric approaches. Based on two pairs of orthogonal linear polarizations with a difference in frequency, which constructs a common-path optical heterodyne interferometer and provides an amplitude-radio mode detection of P polarized heterodyne signal simultaneously, thus the novel PSPWB based on amplitude-sensitive detection method unrolls a high detection sensitivity and a wide dynamic range on effective refractive index at the same time. These advantages will open a window ready for use in measuring the kinetics of low abundance protein-protein interaction in the area such as new drug discovery and highly sensitive protein chip development in the future.