1. Introduction

Surface Plasmon Resonance (SPR) has become a leading technology of choice for studies of biological binding events in real time [1]. SPR biosensors are normally implemented in the Kretschmann-Raether arrangement to pass P-polarized light through a glass prism and reflect it from a gold film (~50 nm), deposited on its surface. SPR effect leads to a dip in the reflected intensity at a defined angle (θ_spr), whose value is resonantly dependent on the refractive index (RI) of a thin 200-300 nm layer near gold. Since most biological species (proteins, DNA etc.) have higher RI than water, the method enables to follow changes of biolayer thickness on gold by a simple monitoring of the resonant angle θ_spr [2] or wavelength [3]. All these make possible the real-time detection of binding events and a fast determination of kinetics constants [4,5,6], which is difficult or even impossible with classical fluorescent label-based technologies [6].

Several recent studies reported improvements of characteristics of SPR biosensors taking advantage of phase-polarization properties of light reflected under SPR. In particular, we showed that the use of phase of light as the sensing parameter instead of the resonant angle or wavelength can lead to a significant upgrade for the detection limit of the method [7,8]. The main idea behind this approach is based on the fact that phase of the p-polarized component experience drastic shift under SPR, whereas the relevant parameter remains almost unchanged for the s-polarized one. The phase shift can then be sensitively measured in various interferometry [7–17] or polarimetry [18] schemes to provide potentially two orders of magnitude gain in sensitivity. Apart from the sensitivity, phase-polarization properties can be used to improve the contrast of the information pattern in conventional SPR [19,20] and to decrease or remove noises in SPR imaging [21].

Providing many undisputable advantages for sensing, phase-polarization SPR methods are often much more demanding in terms of required instrumentation. In particular, to fully take advantage of the high phase sensitivity during a multi-step biosensing experiment, it is necessary to constantly monitor angular or spectral dependence of phase in order to follow the point of the maximum phase response. Generally speaking, such monitoring must inevitably complicate the implementation of compact, low-cost sensor designs, as well as the introduction of feedback loops and multi-channel arrays.

In this paper, we reveal one more remarkable property of phase, related to a weak dependence of the point of its jump on the thickness of the SPR-supporting gold. We show that under a proper sensor design this property can be employed to rapidly determine the optimal zero calibration point and thus adjust the system for maximal phase response. The proposed methodology opens opportunities for the implementation of low-cost phase-sensitive sensor designs for field and multi-sensing applications.

2. Basic idea and approach

First, using Fresnel’s formulae, we performed a numerical simulation of SPR parameters in the Kretschmann-Raether geometry using BK7 glass prism covered by a gold film. It is implied that the gold film contacts either gaseous (n = 1) or aqueous (n=1.33) medium, corresponding to conditions of gas sensing or biosensing, respectively. Fig.1 presents angular dependences for intensity and phase when the thickness of the SPR-supporting gold film is varied (25, 35, 48, 51, 100 nm). One can see that the SPR production leads to the appearance of minima in angular reflectivity curves and jumps in angular dependences for phase. The closer is the thickness to optimal conditions, sharper and more promising are SPR characteristics for sensing. In our system, this optimum was achieved at the gold thickness of 48.8 nm, which provided the smallest reflected intensity at the SPR dip and the sharpest phase jump. It is interesting that the exit of the system from such optimal thickness condition has rather different impact on intensity and phase characteristics. Indeed, as shown in Fig. 1, a variation in the thickness causes a significant shift of the angular position of the SPR dip, whereas the position of the phase jump remains almost the same. The latter fact is clearly illustrated by the existence of a common intersection point for angular phase curves, which appears to be exactly in the point of minimum of intensity for the optimal film thickness, that is in the point of maximal phase sensitivity [7,8]. Thus, the position of the intensity dip is much more sensitive, compared to phase, to the deviation of the gold thickness from the optimal one. We reason that this remarkable property can be explained by different natures of intensity and phase responses. By definition, the intensity of light reflected from an interface is connected to the magnitude of electric field vector, while phase is related to rotations of this vector. It is known that a deviation of the thickness of the SPR-supporting film from the optimal one leads to the increase of losses, which mostly affects the amplitude of electric vector [22]. This leads to a slight variation of photon/plasmon coupling conditions, yielding to a substantial angular shift of the dip position. In contrast, since phase of light is not directly connected to these coupling conditions, it is not so strongly affected by the increase of losses in the system. This leads to a relative independence of the phase jump position of the film thickness.

 

Fig. 1. Angular phase (solid curves) and intensity (dashed curves) curves for different gold layer thicknesses. The data are given for air ambient medium

Download Full Size | PDF

Fig. 2 illustrates how SPR parameters (position of the dip in angular interrogation scheme, the level of light intensity at a fixed angle in intensity interrogation scheme) vary as the thickness of SPR-supporting gold goes out of the optimal conditions (48.8 nm and 48.1 nm for air and water environments, respectively). Notice that the intensity minima are achieved at different angles of light incidence of 43.8 deg. and 72.3 deg., for air and water environments, respectively. One can see that for both environments, the angular position of the resonant SPR dip and the intensity at the fixed angle change significantly (1.2 deg. and 60%, respectively), when the gold thickness varies from 25 to 60 nm. In contrast, the point of intersection of phase curves for different gold thicknesses experiences only a very moderate variation less than 0.07deg. (Fig. 1).

 

Fig. 2. Angular and intensity dependence for the minimum of SPR intensity curves (resonance point) on the SPR-supporting film thickness. Solid and dashed curves correspond to air and water ambiences, respectively

Download Full Size | PDF

Figure 3(a) shows how the phase intersection point, as well as the position of SPR dip, depends on the refractive index of the tested medium. The change of the ambient refractive index from 1 to 1.33, caused by the replacement of air by water, leads to huge angular shifts of both the intensity dip and the phase jump. Nevertheless, it is easy to find that the point of phase intersection always corresponds to the very minimum of intensity in the SPR dip, where the phase sensitivity is maximal [7]. In other words, it means that monitoring the point of phase intersection, one can, with a good precision, follow the position of the SPR dip. It is interesting that a similar phase behavior takes place when the wavelength of light is changed instead of the ambient refractive index (Fig. 3(b)). In fact, it means that similar effects must take place in schemes with spectral interrogation of light when spectral phase is used as the main signal parameter.

 

Fig. 3. Typical SPR intensity (dashed curves) and phase (solid curves) curves for 25nm and 45nm gold film layer (A) Step-like RI change; (B) Wavelength change

Download Full Size | PDF

Summarizing results of theoretical analysis, phase of light reflected under SPR has a remarkable feature related to a relative independence of angular position of the phase jump of th SPR-supporting metal film thickness. Therefore, if two or more films with different thicknesses are deposited on the same sensing slide, there is always a common angle of the beam incidence, which provides the same phase value for all thicknesses. Since such angle corresponds to the reflected intensity minimum, it must provide the maximal phase response. We propose that such phase property can be employed to adjust the biosensor system for maximal phase sensitivity. In the next section, we present an instrumental methodology on the basis of spatial phase modulation with a subsequent Fourier-Transform analysis of the modulated signal to rapidly find the optimal calibration point corresponding to a maximal phase sensitivity.

3. Instrumental methodology

Figure 4 illustrates a schematic diagram of the experimental setup. A 30mW He-Ne laser operating at a wavelength of 632.8 nm is used as the light source. A polarizer and a half-wavelength plate enable to set a linear light polarization with a variable contribution of s- and p-components. The beam is then filtered by an optical filter, consisting of a lens and a pinhole. Changes of light polarization state in the proposed scheme can be considered using Jones transformation matrix method. Here we imply that before spatial modulation, light is linearly polarized with a p-component over the plane XZ and an s-component over the YZ Plane: $E_{\mathrm{INI}}=\left(\begin{array}{c}{E}_x\\ E_y\end{array}\right).$

 

Fig. 4. Schematics of an SPR polarimetry setup for the determination of the zero calibration point

Download Full Size | PDF

The beam is then directed through a birefringent wedge (spatial phase retarder), which spatially modulates phase of the beam. The axis of the retarder is oriented perpendicularly to the plane of the SPR production. After passing the wedge, the beam becomes spatially modulated along the wedge axis with periodic changes of phase relations between the s- and the p-polarized components, giving rise to periodical elliptical polarization (Inset to Fig. 4). This can be described by the following Jones transformation matrix:

where δ is the phase difference between the extraordinary δ_e and ordinary δ_o waves. The birefringent wedge used in this work was made of crystalline quartz with a cut angle of 2.4 deg. The optical axis of the quartz wedge lies in the wedge plane and the refractive indices n_o, and n_e for ordinary and extraordinary beams are 1.5426 and 1.5516, respectively. A half-wavelength plate is placed after the birefringent wedge to compensate additional polarization rotations produced by the wedge and to eliminate SPR polarization fringes. The SPR coupling at a specific incident angle of light on the sensor head (SPR angle) makes possible the excitation of surface plasmons over gold/adjacent medium interface. Since s-polarized light has no phase or amplitude changes under SPR, while p-polarized light experiences changes in both the amplitude A _x and phase, the Jones matrix after the coupling block could be written as:

where ϕ_p and ϕ_s are initial phase components over p and s polarization, respectively, and α is the phase retardation under SPR. An exit polarizer (analyzer), placed just after the SPR coupling block, serves to extract sensor phase information. Light passing through the polarizer with the rotation angle of θ is described by the following matrix:

The analyzer in this work is oriented at θ = 45 degrees to provide interference between projections of the s- and p-polarized components over its axis. The interference results in the appearance of dark and bright fringes associated with the above-stated periodic spatial phase modulation in the wedge. Finally, the spatial distribution of the reflected light intensity is recorded by a CCD camera and examined by an appropriate software image treatment.

The result of such transformations and modulations can be seen in the exit of the system as a spatial distribution of interference fringes, while phase information can be extracted from the position of these fringes. Fig. 5 shows interference fringes for 3 different angular points in the reflectivity curve: (I) before; (II) inside; and (III) after the plasmon resonance.

 

Fig. 5. (A) Typical phase and intensity SPR angular curves; (b) Calculated intensity fringes for several angular positions: (I) before; (II) inside; and (III) after the plasmon resonance.

Download Full Size | PDF

One can see that the fringe contrast is different in these points with the contrast minimum in resonant SPR conditions due to a strong plasmon-related absorption of radiation. However, it is clear that such a decrease of the contrast cannot affect the phase sensitivity, since in the proposed scheme phase information is related to the position of fringes and not to their intensity. As a result, applying methods of Fourier Transform analysis, we can follow the position of fringes and thus extract phase responses to refractive index changes of the ambient medium. One of main advantages of the proposed scheme consists in its relative simplicity. Indeed, in the proposed methodology one can completely avoid the use of costly lock-ins and extract the phase response with a help of a simple photodetector to follow the fundamental harmonics (H_f) of the modulation. This paves the way for a straightforward sensor miniaturization and a cost reduction.

The proposed methodology can be implemented in both imaging and biosensing geometries. In the imaging geometry, SPR is produced in parallel light beam and a typical interference pattern in the exit of CCD camera is shown in Fig. 6(a). In principle, this geometry makes possible the imaging of various objects on gold, as well as a multi-channel sensing [9]. The biosensing geometry does not allow spatial resolution, but it provides sensing measurements in a wide dynamic range. In this case, SPR is produced by a converging pumping beam, while the phase-related fringe pattern is extracted perpendicular to the angular sweep pattern through an appropriate orientation of the wedge axis. A typical interference pattern in the biosensing geometry is shown in Fig. 6(b). Here, the phase variation is significant only within the SPR-related dip (horizontal dark line in the central part of the pattern) that is illustrated by bending fringes in this area. In both cases, the application of the Fourier Transform method enables to fully restore phase information over the laser spot size area. A huge advantage of this method consists in a possibility of efficient image filtering to remove noises and speckles.

 

Fig. 6. Typical images of interference patterns produced in imaging (A) and biosensing (B) geometries

Download Full Size | PDF

4. Results and discussion

To test the proposed methodology and related instrumentation, we carried out experiments in conventional Kretschmann-Raether prism arrangement. The SPR prism coupling system (i.e. the sensor head) used a BK7 glass prism with a proper base angle to provide a near-normal angle of beam incidence to the prism surface. A specially prepared glass slide covered with a thin layer of gold or silver of different thicknesses was placed in immersion contact with the prism. The gold thin films were deposited using E-Beam evaporation on the glass slides or directly onto a BK7 prism. A 3 nm Cr or Ti layer was used to improve the adhesion between gold film and glass substrate.

Fig. 7 shows three different geometries employed for this experiment and CCD images of periodic fringes produced in condition when the parallel light beam is directed on the tested SPR sensor. In the first geometry, three separated gold film stripes with thicknesses of 25, 35 and 51 nm were deposited on a BK7 slide, as shown in Fig. 7(a). The corresponding distribution of fringes from the 3 stripes is shown below the same image. In the second geometry, we used 2 concentric circular areas of different thicknesses of 25 nm and 35 nm (central part), as shown in Fig. 7(b). Finally, the third geometry was a periodic structure with 25 and 35 nm stripes. It has to be noted that in images of interference fringes of Fig. 7 we deliberately selected a non-resonant angle of beam incidence to clearly illustrate the initial phase difference between areas of different thicknesses. This phase difference could be compensated by an appropriate angle, as shown in Fig. 7(c). After a numerical removal of interference pattern noises, we could roughly estimate the initial position of fringes. Then, Fourier-Transform methods were applied to precisely follow the position of interference fringes from areas of different thicknesses. The resultant sensing parameter (PHASE) was an electronic signal describing the position of fringes in all examined areas.

 

Fig. 7. Typical geometries of variable thickness experiment and corresponding distribution of interference fringes over different areas. The angle of light incidence does not match the phase intersection point to clearly show the phase jump between areas of different thicknesses

Download Full Size | PDF

In the calibration experiment, the sensing block was angularly scanned, while phase information from areas of different thicknesses was monitored. Fig. 8 shows results of such experiment for three different gold thicknesses in the geometry of Fig. 7(a). One can see that there was indeed a point of phase intersection, corresponding to the minimum of reflected intensity in the SPR dip under the optimal film thickness (in phase pattern of Fig. 7 this point corresponds to a complete matching of fringe positions for areas of different thicknesses). Notice that the phase curve corresponding to the 51 nm stripe experienced a phase shift due to a change of phase polarity after passing the optimal gold film thickness (this phenomenon is well illustrated in Fig. 1). However, it is easy to find that the angular position of the phase jump for 51 nm stripe was still at the intersection point for other thicknesses. In fact, the addition of an appropriately selected constant to the phase curve associated with the 51 nm stripe set it to the common point of phase intersection.

 

Fig. 8. Experimental SPR intensity (dashed) and phase (solid) curves

Download Full Size | PDF

Thus, using a simple scheme with gold film stripes of different thickness, we were able to easily determine the zero calibration point by a simple angular scanning of the SPR block. We reason that the proposed methodology can be used to simplify the adjustment procedure in real biosensing tests and to decrease the background noise. Indeed, the determination of the intersection point gives the minimum reflectivity point under SPR, which corresponds to the phase sensitivity maximum. In real biosensing experiments, the point of maximal phase sensitivity can change its position after several steps. However, this point can be easily found by angular rescanning of the system. The whole biosensing procedure implies the following steps: 1) determination of incident angle which gives the phase intersection point corresponding to the maximum phase sensitivity; 2) phase measurement using the incident angle; 3) automatic recalibration of the system by fine angle adjustment at the new calibration point after the bio-chemical process is completed. The repetition of the calibrating procedures provides high phase sensitivity and high dynamic range. Other potential advantages of this scheme include the possibility of spatial resolution over the SPR-supporting gold and, as a consequence, the multi-channel sensing in compact designs. Finally, the fact that the phase calibration point coincides with the optimal SPR angle position (minimum intensity of p-polarized light) can be used to improve the dynamic range of phase measurements through the application of back-loop angle adjustments. The latter option is expected to keep the system in conditions of ultra-sensitive phase response in a wide range of RI variations.

5. Conclusions

We introduced a novel methodology and instrumentation to rapidly determine the zero calibration point in phase-sensitive SPR measurements. The application of the sensor chips with different thicknesses of the gold layer thin films enables to implement a reference point in compact designs that greatly facilitates reproducibility and reliability of the phase measurements in field and multi-channel sensing, as well as contributes to the development of feedback loops to improve the dynamic range of measurements.