JP2009019893A - Sensing method and sensing device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve the accuracy of sensing using surface plasmon resonance or localized surface plasmon resonance. <P>SOLUTION: Collimated output light from a light source 101 is condensed onto a plasmon element 105 via a parabolic mirror 104, and reflected light from the plasmon element 105 is collimated again by the parabolic mirror 104. The collimated light from the parabolic mirror 104 is processed by a slit 106, which parallel to the incident surface into the plasmon element 105 and a diffraction grating 107 for imparting dispersion in a direction vertical to the slit 106, and is detected by a two-dimensional detector 108. In an image detected by the two-dimensional detector 108, the dependence on the reflected light by the plasmon element 105 with respect to the wavenumber of incident light into the plasmon element 105 appears in a direction parallel to the incident surface, and dependence with respect to the frequency appears in the direction vertical to the incident surface, and each dependence is analyzed vectorially. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a sensing method and a sensing apparatus using a plasmon sensor used for monitoring changes in surface conditions such as refractive index and temperature, particularly changes in surface conditions caused by antigen-antibody reaction.

  Conventionally, high-sensitivity measurement has been performed by a plasmon sensor that measures wavelength dependency, wave number dependency (that is, incident angle dependency) of plasmon resonance, or both. The dispersion relation of planar surface plasmon resonance (SPR) is

Is given approximately. Here, ε ₁ and ε _m are the dielectric constants of the dielectric and metal that form the transducer interface, and λ is the wavelength. When the wave number of incident light at the transducer interface becomes equal to k _sp (phase matching), the energy of the incident light is combined with plasmons, and the loss associated with the propagation of surface plasmons is observed as the loss of reflected light. The spatial exudation of planar surface plasmons is determined by the dielectric constant of the material that forms the interface between the dielectric and the metal, and is in the order of several hundred nm or less in the visible range. This determines the spatial range with transducer sensitivity.

  On the other hand, in a plasmon interaction in a two-dimensional periodic metal structure, a plasmon electromagnetic field is localized by a boundary condition in a metal unit cell, and a local surface plasmon (LSPR) is induced. The phase matching condition is expressed in a first-order approximation as follows.

Here, m and n are integers, and 2π / Λ is the frequency of the Fourier principal component of the given structure. The LSPR electromagnetic field leakage can be made smaller than that of SPR by localization, and is expected as a means to realize a sensor with higher sensitivity near the surface as a transducer.
JP2003-177089 Review of Scientific Instruments Vol.67, pp.3039-3043 (1998) Physical Review Letters Vol.92, 107401 (2004)

  A conventional SPR sensor detects a change in phase matching condition of SPR due to adsorption or reaction at a metal interface. Since the reflectance change ΔR is measured by observing the signal intensity change ΔI, the minimum detectability is ultimately limited by shot noise.

  Here, e is the elementary charge, P is the incident light power, η is the quantum efficiency of the photodetector, Δf is the detection bandwidth, and hν is the photon energy. That is, at the shot noise limit, the minimum detectability of the reflectance change is proportional to the square root of the incident light power. However, since the plasmon element is a lossy system, when the incident light power is increased, the temperature of the element itself rises due to metal absorption. Therefore, it is desirable to be able to distinguish the shift of the dispersion curve caused by this from the response of the sensor due to the reaction to be monitored.

  In addition, in many conventional SPR / LSPR sensors, the wavelength is fixed and the incident angle is swept, so that measurement time is required and the measurement accuracy is limited by the operation accuracy of mechanical element parts.

  Furthermore, as described above, many of the conventional techniques measure one of the wavelength and the angle spectrum, and thus there is a limit to the information obtained from the transducer. That is, a component due to a disturbance element such as an environmental temperature change of the transducer cannot be separated from a component as a sensor response.

  An object of the present invention is to provide a sensing method and a sensing device with higher accuracy using surface plasmon resonance or localized surface plasmon resonance.

  In order to achieve the above object, the sensing method of the present invention is a sensoring method using surface plasmon resonance in a plasmon element in which a sample is arranged on a metal thin film or a two-dimensional metal periodic structure, or localized surface plasmon resonance. Measuring the change in the dependency of the plasmon element on the wave number and the wavelength of the incident light to the plasmon element, the change in the dependency on the wave number and the wavelength, And a step of analyzing as a shift represented by a vector whose component is an amount related to dependency on wavelength.

  According to the present invention, a sensor response due to a reaction to be monitored and a response due to disturbance are distinguished by an analysis using a vector whose components are a dependency on the wavenumber and a dependency on the wavelength. can do. Thereby, highly accurate sensing is possible.

  FIG. 1 and FIG. 2 (from Non-Patent Document 2) show the dispersion (wavelength and wave number) relationship between SPR and LSPR corresponding to equations (1) and (2), respectively. In the portion indicated by the dark portion, the relationship between the wavelength and the wave number where the incident light and the plasmon are easily phase-matched, and the dispersion curve is reflected. The phase matching condition of SPR has a relationship in which the wavelength and the wave number are inversely proportional, but the phase matching condition of LSPR is defined in the Brillouin zone because it has a periodic structure. What is important is that the dispersion curve of LSPR is more complex than that of SPR and there are multiple modes. Also, this dispersion curve strongly depends on the individual shape of the periodic structure component.

  In a plasmon element in which a sample is arranged on a metal thin film (SPR) or on a two-dimensional metal periodic structure (LSPR), a sample change (change in surface state) due to a reaction to be actually monitored locally occurs. Consider the adsorption of a substance at the metal structure interface. When the amount of change in refractive index and thickness is Δd and Δl, respectively, the effective amount of change Δs is

Call it. The change in reflectance ΔR under fixed measurement conditions is defined as θ, the incident angle of incident light to the plasmon element, and λ,

Given in. In order to increase the sensitivity of the sensor, focus on θ _1/2 and λ _1/2 that maximize the variable differential term with respect to the reflectance R, and observe so that the reflectance change rate A is maximized. desirable. As a result, sensing can be performed with the highest performance. Generally used SPR sensors use only the first or second term in the square root (projection to one component in Figs. 1 and 2). In addition, the techniques disclosed in Patent Document 1 and Non-Patent Document 1 can measure the dependence of reflectance on θ and λ, but do not disclose a technique for sensing by paying attention to the interpretation of the above formula. From the equation (5), the sensor response vector is defined as follows.

  Here, it is shown that the sensor response vector p is different from an arbitrary disturbance response vector q. As an example, assume that a change in the temperature of the element itself has occurred, thereby causing a disturbance. The disturbance response vector in this case depends on the refractive index of the metal and the temperature change of absorption, but these can be replaced by the change in dielectric constant, and the resulting change in reflectance is

It becomes. The resulting band image change

It becomes. This is generally different in direction from the vector p, so that temperature changes and sensor responses can be distinguished. As a matter of course, this effect becomes larger as the wavelength is closer to the plasma frequency of the metal. If measurement is performed so that the region where the sensor response vector p is large and the region where the disturbance response vector q is large are included in the same image, it is easy to distinguish the effect of the disturbance. Here, the disturbance element is illustrated as a temperature change, but even if a plurality of disturbance elements are included, the image change is reflected as a linear sum of them. When the source of the disturbance is clearly known, it is possible to define a new disturbance element vector and extract the vector element.

  In the present embodiment, the description will be made with attention paid to the reflectance R of the plasmon element, but the same argument can be made even if attention is paid to the transmittance.

(Example 1)
In the sensing method of this embodiment, a plasmon band image is measured with an apparatus that eliminates the mechanical sweep mechanism of wavelength and angle, and analysis is performed on the vector shift of the band image. For measurement, the following methods can be considered for normal incidence and oblique incidence. However, in the case of an LSPR sensor, considering that it is only necessary to examine the Brillouin zone, direct incidence using high NA incident light is considered sufficient, but the sensitivity of the sensor response at the band edge is good, and the information When it is desired to close up, oblique incidence is also conceivable.

  3 and 4 show an example of a sensing device that executes the sensing method of the present embodiment.

  In the case of normal incidence, as shown in FIG. 3, the output light from the light source 100 that outputs collimated light is adjusted to a polarization axis with a good response by a polarizer (or wave plate: depending on the nature of the light source) 102, and a half mirror After passing through 103, the light is condensed on the plasmon element 105 using the parabolic mirror 104. The reflected light from the plasmon element 105 is collimated again by the parabolic mirror 104, and the component reflected by the half mirror 103 passes through the slit 106 parallel to the incident surface to the plasmon element 105. The two-dimensional detector 108 detects an image generated through a diffraction grating 107 that gives dispersion in a direction perpendicular to the slit 106.

  In the case of oblique incidence, as shown in FIG. 4, the output light from the point light source 201 passes through a polarizer (or wave plate: depending on the nature of the light source) 202, and then condenses using an ellipsoidal mirror 203 on the plasmon element 204. Then re-image. The reflected light is collimated by a parabolic mirror 205, and an image generated by passing through a slit 206 parallel to the incident surface of the plasmon element and a diffraction grating 207 that gives dispersion in a direction perpendicular thereto is displayed as a two-dimensional detector 208. Detect with.

  With these sensing devices, dependency on the incident angle in the direction parallel to the incident surface to the plasmon elements 104 and 204 (thus, dependency on the wave number k), and dependency on the wavelength λ in the direction perpendicular to the reflection surface and the incident surface. Mapping (see Figure 5). That is, in the image detected by the two-dimensional detector 208, a change in reflected light appears when the incident angle is changed in a direction parallel to the incident surface to the plasmon elements 104 and 204, and a direction perpendicular to the incident surface. Then, a change in reflected light appears when the wavelength is changed. Therefore, the dependence on the incident angle, that is, the wave number and the dependence on the wavelength can be obtained collectively by analyzing the obtained two-dimensional image without using a mechanical working part. That is, the band image of the plasmon element can be measured at high speed, and the operation accuracy of the mechanically operating portion does not affect the measurement accuracy.

Here is a summary of important points.
a) A point light source image is adjusted to the required polarization and re-imaged on the plasmon element for batch measurement without angular scanning, and the reflected light from the plasmon element is collimated, parallel to the incident surface. Dependence on the incident angle appears.
b) A two-dimensional detector that gives angular dispersion in the direction perpendicular to the incident surface by a slit parallel to the incident surface and a diffraction grating arranged so that the groove is parallel to the incident surface with respect to the collimated reflected light. Image on top.

  In FIGS. 3 and 4, simplified diagrams are shown to clarify the functions, but various additional functions can be added in order to enhance the functions of the respective functions. For example, the light source may be a halogen lamp-based fiber light source, or a supercontinuum light source to suppress excessive heating of the element. However, in the latter case, the collimated light may be reduced by a parabolic mirror instead of the ellipsoidal mirror. In Patent Document 1, a lens system is specified for collective measurement of angles, but here, a reflection optical system is used in consideration of obtaining a band image in a wide wavelength range. However, a lens system may be used if the wavelength range of the band image is narrow. Regarding the slit, the reflected light from the plasmon element may only be imaged on the entrance slit of the imaging spectrometer. In addition, if it is desired to close up a specific place in the band image with high resolution, an optical element that collimates the wavelength dispersion direction may be inserted before inserting the magnifying optical system (for example, the lens system in FIGS. 109,209). In LSPR, the group velocity is effectively reduced at the band center and band edge, so that a larger interaction can be obtained. That is, depending on the design of the element, there is a case where the sensor sensitivity is high at the band edge and the oblique incidence optical system is effective.

  Next, the band image obtained using the above-described apparatus is analyzed. The first procedure is to identify a high-sensitivity region in a two-dimensional space whose components are wave number and wavelength. Next, as a second procedure, the sensor response is determined by paying attention to the identified high sensitivity region. The sum of the brightness of the image in the high-sensitivity region can be simply used as the sensor response, but a predetermined condition is imposed on each image element, image processing is performed accordingly, and the sensor response vector p and disturbance vector q are It is also possible to extract. Some examples are shown below.

(Identification of high sensitivity position)
First, in a state where there is no sensor response, the reference image R _ij (i = 1 to M, j = 1 to N) is measured. This image is subjected to partial differentiation with respect to k and λ (D ^k _ij and Dλ _ij ; see FIG. 6). Next, N shift vectors S ^k _j can be obtained by evaluating the extreme values of the slice image of the reference image R _ij and sensor response image F _ij for a certain i and _calculating the shift amount in the k direction. (See Figure 7.) When the same operation is performed on a certain j, M shift vectors Sλ _{i in the} λ direction are obtained. The extreme value of the partial differential image may be used for the calculation of the shift amount. It is important to be able to determine whether the distribution of the shift vector is significantly changed, so that the sensor response is non-linear or another factor is included. A k-direction shift matrix (S ^k _ij ) having the same column elements as the k-direction shift vector and a λ-direction shift matrix (Sλ _ij ) having the same row elements as the λ-direction shift vector are defined, and the following image is created.

This corresponds to equation (5) under the assumption that the shift occurred uniformly. That is, the region where P _ij is maximized corresponds to the region where A in Equation (5) is maximized, and gives a highly sensitive region. FIG. 8 is an image generated by equation (7). In this case, P _ij has a maximum value at a wavelength near 1 μm, and it can be seen that the sensor response is relatively sensitive. In equation (7), if the differential image (G _ij = R _ij -F _ij ) between the reference image and the sensor response image is used instead of the partial differential image D _ij , this includes only total differential information. It does not directly correspond to equation (5).

By the way, considering the identification of the high sensitivity region using G _ij as a simpler processing method,

A highly sensitive region may be identified. An example of an image obtained by this operation is shown in FIG. In this case, the wavelength at which P _ij takes the maximum value is around 900 nm. However, in this case, since the total differential term and the variable differential term are not independent, there is a possibility that a portion where the sensitivity is artificially increased is generated as a result of multiplication of orthogonal elements. This is because even if the shift to either k or λ is very small, if the profile in that direction is sharp, it becomes a highly sensitive part.

Returning to this discussion, after the high sensitivity region is identified, the difference intensity at the point where P _ij takes the maximum value may be used as an index of the sensor response. Alternatively, a certain range (i, j: P _ij > P _c : const) may be defined, and a value obtained by integrating G _ij in the range may be calculated as an absolute response. In the latter case, the effect of noise can be reduced in inverse proportion to the square root of the number of elements.

(Sensor response discrimination after element extraction)
In a range that can be captured as a projection in a linear operation, a polynomial approximation can be performed on the extracted dispersion curve (ridge line formed from peak points) to obtain the size of the projection vector. For example, the dispersion curves extracted from the reference image and sensor response image are

When we can approximate

Define the sensor response vector as

Can also be used as a sensor response. It is possible to determine the possibility that the sensor response deviating from this vector includes a disturbance. In this case, if N data points are used in the approximation of equation (9), the noise can be reduced by √N times, and high sensitivity can be achieved by applying to a large number of sample points. For approximation of equation (9), data points outside the above-described high sensitivity region range can also be used. If this is similarly applied to a partial differential image, a plurality of dispersion curves can be defined for one mode in a pseudo manner, so that further noise reduction can be achieved.

(Vector element extraction and decomposition from differential peak curve)
In the sensor response determination described above, a plurality of elements of the sensor response vector are separated into linearly independent elements, but the constituent elements are not identified. In the following, as a more general analysis method, consider extracting orthogonal elements from a difference image. This is a unit vector for sensor response (elements are integers)

And the unit vector for the disturbance response

, The difference image component is considered to be the total derivative with respect to the perturbation, so

It becomes. Taking the same difference image with the sample concentration a times, for example,

Is obtained. Taking a difference image in the same way after changing the temperature as a disturbance element, for example,

Get. From the above ΔS ₁ and ΔS ₂

Therefore, it can be reduced to an algebraic equation including 11 variables of dk _r , dk _p , dλ _r , dλ _p , β ₁ , β ₂ , β _T , α ₁ , α ₂ , κ ₁ , κ ₂ . Since there are only four variable parameters on the dispersion surface, dk _r , dk _p , dλ _r , and dλ _p , that we want to know here, along the dispersion curve (the curve formed by the extreme values of the difference image) (k _i , λ _{For l} ) (l = 1 ... N), 4N + 7 points should be selected from the neighborhood. In the case of LSPR, attention may be paid to the dispersion curve of the most sensitive mode (the norm of the sensor response vector p is large), or the above equation may be applied to n modes. An appropriate range should be selected according to the resolution of the image and the SNR of the signal. If a calibration curve map is created in advance by changing several values of a, the sensor response vector p is nonlinear with respect to the concentration because the position of the dispersion curve and the vector direction can be fitted simultaneously. Can also respond. Fitting is a linear sum for each pixel

And fit f to the difference between the sensor response image and the reference image (Fig. 10) (see Fig. 11). FIG. 10 is an example of a difference image, where the left is a pure sensor response and the right is a response accompanying a change in the dielectric constant of the metal. Referring to FIG. 10, it can be seen that there is a local maximum (white) and minimum (black), and in the response accompanying the change in the dielectric constant of the metal, the change is biased toward the short wavelength side.

A pair of a sensor response vector p and a disturbance response vector q that minimizes the fitting error is selected, and X _{1 corresponding} thereto is taken as a sensor response. If the fitting error does not fall within a certain standard deviation, it can be measured again as an error. Furthermore, even when the sensor response vector p is smaller than the disturbance vector, if the time change of the disturbance response vector q is isotropic on the kl plane, the M A sum may be taken for the series. Accordingly, the sensor response can be extracted by utilizing the fact that the disturbance response vector q is canceled and only the sensor response vector p remains.

  In the above explanation, it is assumed that the sensor response vector p and the disturbance response vector q are included in the image one element at a time. However, it is possible to process a band image using various assumptions other than those described above, for example, a plurality of disturbance elements are included.

An example of wavelength and wave number dependent reflectivity reflecting the band structure of a planar surface plasmon (vertical axis: wavelength λ, horizontal axis: wave number k). An example of wavelength and wavenumber dependent reflectivity reflecting the band structure of localized planar surface plasmons. The schematic diagram which shows the sensing apparatus of an example in the case of direct incidence. The schematic diagram which shows an example of a sensing apparatus in the case of an oblique incidence. Conceptual image just before the slit and on the 2D detector surface. Variable differential image in k direction and λ direction (vertical axis: wavelength λ, horizontal axis: wave number k). The conceptual diagram of shift vector extraction. An image generated by equation (6) (vertical axis: wavelength λ, horizontal axis: wave number k). An image generated by the processing according to equation (7) (vertical axis: wavelength λ, horizontal axis: wave number k). Example of difference image. Vector function sum expression corresponding to each pixel of the difference image.

Explanation of symbols

101 light source
102 Polarizer (or wave plate)
103 half mirror
104 Parabolic mirror
105 Plasmon element
106 slits
107 diffraction grating
108 2D detector
201 Point light source
202 Polarizer (or wave plate)
203 Ellipsoidal mirror
204 Plasmon element
205 parabolic mirror
206 Slit
207 diffraction grating
208 2D detector

Claims (5)

In a sensing method using surface plasmon resonance in a plasmon element in which a sample is arranged on a metal thin film or a two-dimensional metal periodic structure, or localized surface plasmon resonance,
Measuring a change in dependence of the reflectance or transmittance of the plasmon element on the wave number and wavelength of light incident on the plasmon element;
Analyzing the change in the dependence on the wave number and the wavelength as a shift represented by a vector whose components are the quantity relating to the dependence on the wave number and the quantity relating to the dependence on the wavelength;
A sensing method characterized by comprising:   In a two-dimensional space having wave number and wavelength as components, obtain a region where the change rate of the reflectance or transmittance of the plasmon element is maximum when the dependence on the wave number and wavelength changes, and in that region The sensing method according to claim 1, wherein the vector shift is analyzed.   The sensing method according to claim 1 or 2, wherein shifts of a plurality of peak points of the surface plasmon resonance or the localized surface plasmon resonance are simultaneously measured. A sensing device that executes the sensing method according to any one of claims 1 to 3,
A light source that outputs collimated light;
A polarizer or wave plate through which the output light from the light source is passed;
A parabolic mirror that condenses the light that has passed through the polarizer or wave plate on the plasmon element and collimates the reflected light from the plasmon element;
A half mirror disposed between the polarizer or wave plate and the parabolic mirror;
A slit parallel to the incident surface of the plasmon element, through which light collimated by the parabolic mirror and reflected by the half mirror is passed;
A diffraction grating that provides dispersion in a direction perpendicular to the slit with respect to the light passed through the slit;
A two-dimensional detector for detecting light from the diffraction grating;
A sensing device. A sensing device that executes the sensing method according to any one of claims 1 to 3,
A point light source,
A polarizer or wave plate through which the output light from the point light source passes,
An ellipsoidal mirror for condensing the light passing through the polarizer or wave plate on the plasmon element;
A parabolic mirror for collimating the reflected light from the plasmon element;
A slit parallel to the incident surface of the plasmon element, through which light collimated by the parabolic mirror is passed;
A diffraction grating that provides dispersion in a direction perpendicular to the slit with respect to the light passed through the slit;
A two-dimensional detector for detecting light from the diffraction grating;
A sensing device.