WO1997015819A1 - Surface plasmon resonance light pipe sensor - Google Patents

Abstract

The present invention provides SPR sensors in which the sensing element is a planar light pipe (20). The sensors of this invention include configurations which employ multiwavelength light incident on the SPR sensing area (30) at a single angle or at a range of angles. Sensors of this invention also include configurations that employ monochromatic light at a range of angles. Many of the configurations of the SPR light pipes of this invention involve imaging of input light through the light pipe. In one embodiment, the invention provides a first order SPR sensor system in which the sensing element is a planar light pipe. Light coupled into the light pipe reflects off an SPR sensing area positioned on an external surface planar surface of the light pipe. Multiwavelength light that is coupled into the light pipe input face at a range of angles propagates through the light pipe by total internal reflection (TIR), making multiple reflections, and exits in a series of angular bands each containing spectral information (including SPR features) for a small range of incidence angles. A detector or detectors are positioned to measure the reflection spectrum, including any surface plasmon resonance feature, of one or preferably more than one of the angular bands exiting the light pipe. SPR sensor configurations include those that have multiple sensing channels and those which can be multiplexed. This invention provides SPR sensors with planar light pipe sensing elements and method of detecting analytes in samples using these sensors. The invention also provides planar light pipe that have a plurality of SPR sensing layers on a planar surface.

Description

SURFACE PLASMON RESONANCE UGHT PIPE SENSOR

This invention was made, at least in part, with support from the National Science Foundation under grant number EID-9212314. The United States Government has certain rights in this invention.

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of priority under 35 U.S.C. 119(e) of U.S. provisional applications serial numbers 60/007,027 filed October 25, 1995; 60/005,878 filed October 26, 1995; and 60/009,169 filed December 22, 1995 all of which are incoφorated in their entirety by reference herein.

HELD OF THE INVENTION This invention relates in general to surface plasmon resonance sensors and more particularly to zero and first order sensors employing a planar lightpipe sensor configuration. Configurations of these sensors can be employed for eidier or both wavelength modulation or angular modulation SPR.

BACKGROUND OF THE INVENTION Optical surface plasmon resonance (SPR) sensors are sensitive to changes in the refractive index (RI) of a sample near the sensor surface. Most bulk prism SPR sensor configurations measure either the angular reflection spectrum for monochromatic light or the wavelength reflection spectrum for collimated white light. Various SPR sensor configurations utilizing waveguides, including optical fibers, have been reported. These include a single-mode planar waveguiding structure which detects intensity changes in monochromatic light (Lavers, CR. and Wilkinson, J.S. (1994) "A Waveguide-Coupled Surface-Plasmon Sensor for an Aqueous Environment" Sensors and Actuators B 22:75-81) and a sensor system in which white light is injected into a single-mode waveguide having an SPR supporting superstructure (Kreuwel, H.J.M. et al. (1987) "Surface Plasmon Dispersion and Luminescence Quenching Applied to Planar Waveguide Sensors for the Measurement of Chemical Concentrations," Proc. SPIE 798:218-224; Lambeck, P.V. (1992) "Integrated Opto-Chemical Sensors" Sensors and Actuators B 8:103-116). Additionally, a white light multi-mode fiber optic SPR sensor has been introduced (U.S. Patent 5,359,681, issued October 1994); Jorgenson, R.C. and Yee, S.S. (1993) "A Fiber Optic Chemical Sensor Based on Surface Plasmon Resonance," Sensors and Actuators B 12:213; Jung, C.C. et al. (1995) "Fiber-Optic Surface Plasmon Dispersive Index Sensor for Highly Opaque Samples" Process Control and Quality 7:167- 171; Jorgenson, R.C. and Yee, S.S. (1994) "Control of the Dynamic Range and

Sensitivity of a Surface Plasmon Resonance Based Fiber Optic Sensor" Sensors and Actuators A 43:44-48; Mar, M. et al. (1993 ) "In-Situ Characterization of Multilayered Langmuir-Blodgett Films Using a Surface Plasmon Resonance Fiber Optic Sensor" Proc. of the 15th Annual Conf. of the IEEE Engineering in Medicine and Biology Soc, San Diego, CA pp. 1551-1552.

U.S. Patent 5,485,277 (filed July 26, 1994, issued Jan. 16, 1996) "Surface Plasmon Resonance Sensor and Methods for the Utilization Thereof " reports an SPR sensor said to comprise a "waveguide" cartridge, a cylindrical diverging lens coupled to the "waveguide" and a plurality of photodetectors optically connected to the cylindrical lens. The "waveguide", exemplified by a microscope slide with angled input and output faces, carries a symmetrically positioned metal layer that supports SPR. Apparently, monochromatic light is introduced into the "waveguide" through the angled polished input face by focusing the light through the end of the "waveguide" onto the metal sensor surface at a range of angles spanning angular location of the surface plasmon resonance. The angle of the input face of the "waveguide" is polished to an angle α^ , where tanα = tanθ , - ^{Se dc}"^art (1) wvg

and n^ is the refractive index of the "waveguide" and θ_ceMral is the center angle of the range of rays directed at the sensing surface. It appears that θ_cemral is very close to Θ_SPR , the surface plasmon resonance angle. The RI of the sample is determined by measuring light intensity exiting the sensor as a function of angle. The patent also discusses the use of sensing and reference channels on the same metal film-coated waveguide.

SPR sensor waveguide configurations that have been reported are limited to those measuring refractive index at either a single wavelengdi or a single angle (angular or wavelength modulation, respectively). Such configurations are considered zero order sensors since they measure only one independent variable for a given analyte. Recently, a first order SPR sensor geometry that can simultaneously measure a sample's index of refraction at multiple wavelengd s was reported (Karlsen, S.R. et al. (1995),

"Simultaneous Determination of Refractive Index and Absorbance Spectra of Chemical Samples Using Surface Plasmon Resonance," Sensors and Actuators B 24-25:747-749). This dispersive RI sensor which employed a cylindrical sapphire prism with a gold sensing layer required several discrete optical components which introduced optical aberrations in the reflected signal, making it difficult to calibrate both the angular and spectral outputs of the sensor.

Thus mere remains a need in the art for first order SPR sensors that allow independent measurements of two variables for a given analyte providing dispersive RI information and for SPR sensors allowing the simultaneous determination of two parameters, for example, film thickness and RI of a diin film applied to an SPR sensing surface. There is also a need in me art, particularly for assay of biological samples, for SPR sensors that can simultaneously detect more d an one analyte in a sample (multiplexed sensors). There also generally remains a need for SPR sensors which are sensitive, simple to use, readily calibrated, compact in size and rugged, and inexpensive to produce. SPR sensors of diis invention meet these needs. SUMMARY OF THE INVENTION The present invention provides SPR sensors in which e sensing element is a planar lightpipe. The sensors of diis invention include configurations that employ multi- wavelengm light (including broad band and white light) incident on e SPR sensing area at a single angle or at a range of angles. Sensors of this invention also include configurations that employ monochromatic light at a range of angles. Many of the configurations of the SPR lightpipes of this invention involve imaging of input light dirough the lightpipe.

In one general embodiment, die invention provides a first order SPR sensor system in which the sensing element is a planar lightpipe. Light coupled into the lightpipe reflects off an SPR sensing area on a planar surface of die lightpipe. The lightpipe sensor is distinct from previous waveguide systems in that ray tracing theory describes me propagation of light in and dirough die lightpipe at a range of angles. In contrast, light propagation in waveguides (eimer single- or multiple-mode) is described by mode Uieory. Non-monochromatic, i.e., multiwavelength, light diat is coupled into die lightpipe input face at a range of angles propagates irough me lightpipe by total internal reflection (ITR), making multiple reflections, and exits in a series of angular bands each containing spectral information (including SPR features)for a small range of incidence angles. A detector or detectors are positioned to measure the reflection spectrum, including any surface plasmon resonance feature, of one or preferably more than one of die angular bands exiting the lightpipe. The same sensor configuration can be employed to measure SPR from a range of incidence angles of monochromatic input light (simple angular modulation)or to measure SPR from a range of incidence angles of more dian one wavelengdi. The measurement of a broad range of angles and wavelengtiis allows die measurement of a dispersive RI curve (i.e., RI as a function of wavelengdi) of a given sample or analyte.

The lightpipe SPR sensors of diis invention have automatic angular calibration, minimal optical aberrations, and does not require index matching fluids. In one embodiment, a lightpipe sensor can be fabricated from inexpensive, disposable microscope slide. Providing an SPR sensor which can measure dispersive RI of a sample in a simple, compact, and inexpensive manner.

Light can be coupled into and out of the lightpipe sensors of this invention in a variety of ways using conventional optical apparatus to focus, collimate and/or selectively expand light on coupling into the lightpipe or to focus, collimate, disperse, and/or collect light exiting the lightpipe. Incident light must comprise TM polarized light. The light employed in the sensor is optionally TM polarized to remove die TE component prior to launch into the sensor. Alternatively, me TE component light of light is removed by passage through a TM polarizer any time prior to detection.

The planar lightpipe of me sensors of this invention has at least one SPR sensing area on at least one of its planar surfaces. A SPR sensing area comprises an SPR- supporting conducting layer, preferably an SPR-supporting metal layer and optionally has an adherence layer, dynamic range-controlling layers and reactive layers. In particular embodiments, the SPR sensors of iis invention can have a plurality of sensing areas across the widdi, along die length or bom of a lightpipe surface. Lightpipes of me sensors of this invention can include bom active and reference sensing areas and active sensing areas on a given lightpipe can have sensing areas specific for the same or different analytes in a sample.

SPR sensors of this invention can include static or flow sample cells to confine samples for SPR measurements. This invention includes multichannel and mutiplexed SPR sensors having a plurality of independent SPR sensing areas on a planar lightpipe surface. In addition, sample cells can be configured to contact different sensing areas on the lightpipe surface wim different samples.

Lightpipes of the sensors of diis invention can be fabricated from a variety of materials transparent or semi-transparent to the input light. Glass, crystal, including sapphire, as well as plastic and polymer materials can be used for lightpipe substrate.

Lightpipes optionally have a cladding layer to insure TIR of light along me length of the lightpipe. SPR Lightpipe sensors of diis invention can be employed as biosensors, particularly to detect multiple analytes in biological samples, such as serum or blood.

In a specific embodiment, a planar lightpipe sensor configured for single angle operation is provided. The lightpipe is beveled at selected angles at its input end or at bodi its input and output ends to facilitate coupling of substantially collimated white light, preferably TM polarized white light, at a selected single angle mat excites SPR at me sensing area. Angle selection is made by bevel angle used. This embodiment is a zero order sensor in the sense that it allows measurement for a given analyte at only a single angle of incidence. In diis embodiment, however, the lightpipe can have a plurality of SPR sensing area across its wid i to provide for multichannel sensing. Refractive index sensitivity of this configuration is estimated as 4 x 10'⁵ RI units.

This invention provides SPR sensors wim planar lightpipe sensing elements and method of detecting analytes in samples using these sensors. The invention also provides planar lightpipe that have a plurality of SPR sensing layers on a planar surface.

BRIEF DESCRIPTION OF THE HGURES Figure 1 is a schematic drawing of a prior art bulk prism angle modulated SPR sensor illustrating the Kretschmann configuration.

Figure 2 is a schematic drawing of a prior art bulk prism wavelengdi modulated SPR sensor analogous to die Kretschmann configuration of Figure 1.

Figure 3A is a schematic drawing of a lightpipe SPR sensor of this invention in which input light is end-coupled into die lightpipe at a range of angles below me axis of the lightpipe. Light exits me lightpipe in an angular array of bands alternating above and below the axis.

Figure 3B is a perspective view of the lightpipe sensor of Figure 3A showing the placement of the sensing region on the external top surface of me lightpipe. Figure 3C provides the experimental reflection spectra for bands 5, 7 and 9 from the device of Figure 3 A where water is me sample.

Figure 4 is a drawing that illustrates unfolding a lightpipe multiple times widi die straightened light rays passing through me virtual location of die top surface multiple times. A_{b r} refers to me r"¹ top surface reflection location for maximum angle ray of the b* angular band and B_{b 5} refers to die reflection location for the corresponding minimum angle ray.

Figure 5 is a graph of lightpipe top surface reflections (where the sensor is positioned) as a function of angular band number showing the location and size of illuminated spots on die top surface of die light pipe for each angular band as horizontal solid line segments. A reflection extends from the beginning to end of die line segments. The shaded area is die location of die SPR-supporting metal layer, i.e., the sensing area. Band numbers are as illustrated in Figure 3 A and as indicated in Figure 4.

Figure 6 is a graph of me results of calculations of die number of reflections for each light ray off me metal sensing surface of me lightpipe where the position of the sensing surface is as shown in Figure 5. The calculation indicates mat bands 5, 7, and 9 will experience 1 , 1, and 2 reflections, respectively, while bands 1, 3, and 4 will all only experience partial reflections. The graph shows the range of angles of each band.

Figure 7 is a schematic drawing of a planar lightpipe sensor of this invention in which light is end-coupled into d e lightpipe by focusing light at die input face at a range of angels above and below me lightpipe axis (symmetrical light input). Input light filling angles above the axis exits d e lightpipe in an angular array of bands alternating above and below die axis. This exit pattern is complementary to the exit pattern of bands from input light filling angles below die lightpipe axis.

Figure 8 illustrates an alternate configuration of me asymmetric input SPR sensor of Figure 3A wherein the exit optics of die lightpipe employ a telecentric lens to redirect and focus angular output bands. Figure 9A illustrates an alternative first order SPR sensor configuration with light input into the lightpipe through the lower planar surface and off a mirrored bevel on die input end of d e lightpipe. Light exits die lightpipe dirough a built-in diffraction grating on a beveled out put end of die lightpipe. Figure 9B illustrates an idealized contour plot detected at the output plane by an image detector of die SPR sensor of Figure 9A of die reflection coefficient of die output light as a function of bod wavelengdi and angle.

Figure 10 compares theoretical values of die SP resonance wavelengdi (λ_SPR) and experimental measurements of λ_spr determined using a sensor of Figure 9A. λ_spr is plotted as a function of the center angle for each angular band for samples of acetone and water.

Figure 11 illustrates a particular embodiment of a planar lightpipe zero-order SPR sensor configuration particularly useful for single-angle light input. The lengd of me lightpipe (L), die widdi of die sensing area (W) and the bevel angle ct are selected to maximize RI sensitivity.

Figure 12(a)-12(d) are SPR refection spectra measured with die single-angle SPR lightpipe sensor of Figure 11 for aqueous glycerol solutions of concentrations (weight percent of solution) of (a) 4.02%, (b) 6.22%, (c) 7.30%, (d) 9.50% .

Figure 13 is a graph of SP resonance wavelengdi (λ_SPR) of aqueous glycerol solutions as a function of concentration (weight percent of solution) showing me linear calibration curve for a sensor of Figure 12. The corresponding refractive indices of die solutions are shown on d e alternate axis. Experimental points are indicated by "*" and die solid line indicates curve fitting.

Figures 14A and 14B illustrate a dual-channel configuration of the SPR sensor of Figure 7 wim symmetric light input. Figure 14A is a top view showing die lateral placement of me two sensing areas. Figure 14B is a side view of me sensor showing input and output optics. Figure 15 is a drawing of a dual-channel configuration of me SPR lightpipe sensor of Figure 11.

Figure 16A is a graph of λ_SPR measured at bodi channels of me dual-channel SPR sensor of Figure 14 as a function of glycerol concentration. Figure 16B is a graph in which the derivatives of die concentration curve (δλs_PR/δconc) of Figure 16A are plotted as a function of glycerol concentration to compare the SPR signal shifts due to me refractive index variations in me spectra from two channels representing two different incidence angles on d e sensing surface.

DETAILED DESCRIPTION OF THE INVENTION A surface plasmon wave is an electromagnetic wave which propagates along the interface between a conductor (or semi-conductor) and a dielectric, and decays normal to the interface. A plasmon is the collective oscillations of free charges, ions or valence electrons, in a metal or semi-conductor which can be excited by a polarizing interaction (a polaritron) between an electromagnetic wave and an oscillator resonant at die same frequency as the wave. A surface plasmon polaritron is the interaction between photons and d e collective oscillations of electrons at die surface of a conductor. The interaction is strongest at the resonance condition known as surface plasmon resonance, which is satisfied when the tangential component of d e wave vector of light incident on d e conductor is equal to d e wave vector of me surface plasmon wave. Incident light satisfying d e resonance condition causes surface charges on the conductor to oscillate creating a bound electromagnetic or charge-density wave propagating along die interface between me conductor and a dielectric material. The resonance condition depends on die wavelengdi of incident light and the angle at which light is incident upon die interface as well as die dielectric constants of all of die materials in die layers involved, including diat of a dielectric sample in contact with die sensing layer.

For die case in which a metal SPR-supporting layer on a supporting dielectric substrate (e.g., glass) is in contact with a dielectric sample, the resonance condition is:

*x = substrateSin(θ) = £_5p, (2)

where k₀= 2π/λ is die free space wave vector of die incident light, θ is die incident angle of die light, n_subsmιe is d e complex, wavelengdi-dependent refractive index (RI) of die substrate. The wave vector of surface plasmon wave, k^, can be approximated as (Jung, C. (1991), "Surface Plasmon Resonance," Master Thesis, University of Washington)

where ε_d and ε_m are die wavelength dependent complex dielectric permitivities of the dielectric sample and metal.

In a sensing configuration, die excitation of SPR is detected as a decrease in die reflection coefficient (i.e., a decrease in intensity) of TM polarized light reflected off die sensing interface. See: Pockrand, I. et al. (1979) "Exciton-Surface Plasmon Interactions"

J. Chem. Phys. 70:3401; Jorgenson and Yee (1993) supra. Plane-wave Fresnel reflection equations can be used to model die resonance condition for a sensing interface containing multiple planar layers. See: Ishimaru, A. Electromagnetic Wave Propagation. Radiation. and Scattering. Prentice Hall, Englewood Cliffs, NJ (1991) Chapter 3 pp. 43-45.

Conventional SPR sensing techniques generally use either angle or wavelengdi modulation. These techniques are illustrated in Figures 1 and 2 for a Kretschmann configuration (Kretschmann, E. and Raedier, H. (1968) "Radiative Decay of Non- radiative Surface Plasmons Excited by Light" Z. Naturforsch., Teil A, 23: 2135-2136) in which a d in, highly- reflective metal film (2) is deposited on die base of a prism (3). The metal layer is brought into contact wim a solid, liquid or gas dielectric sample (4) creating a metal-dielectric interface that will support SPR. In the Kretschmann configuration of Figure 1 , TM-polarized monochromatic light is focused onto die back of die sensing surface (metal layer, 2) at a range of angles (5) and die intensity of the light reflected from the sensing surface is measured as a function of incidence angle, for example, using a linear array detector (6). This is angle modulation. Figure 2 illustrates a wavelength modulated SPR sensor analogous to d e Kretschmann configuration of

Figure 1. TM-polarized collimated white light is incident upon die sensing surface (2) at a single angle (7), and intensity of the light reflected from that surface is measured as a function of wavelengdi, for example, using a spectrograph (8). As stated above, resonance depends upon bom the wavelengdi of incident light and die angle or angles at which e light hits die sensing interface (incidence angle(s)). These traditional zero order sensing techniques hold one variable (angle or wavelengdi) constant and measure d e reflection coefficient as a function of the odier variable. For angle modulated SPR at a fixed wavelengdi, mere can be an angle Θ_SPR diat satisfies d e resonance condition. For a fixed angle, diere can be a wavelengdi λ_SPR at which die wavelength dependent permitivities (or RI's) of me various media are such as to satisfy die resonance condition. The angle or wavelength at which resonance (i.e., minimum reflected light intensity) is observed gives a measure of the effective index of refraction (RI) of the dielectric sample. The resonant angle (0^.) or wavelengdi (λ_spr) can be calibrated (using samples of knownrefractive index) to me refractive index of die sample.

Surface plasmon resonance sensors using only wavelength or angle modulation are considered zero order sensors because die vector of reflected intensities (Iθ) or I(λ)) is reduced to a single value of (θ_spr) or (λ^). The previously reported bulk optic first order SPR sensor (Karlsen, S.R. et al. (1995), "Simultaneous determination of refractive index and absorbance spectra of chemical samples using surface plasmon resonance," Sensors and Actuators B lA-lS-.lW-ltø) simultaneously monitored reflection coefficients over a range of angles and a range of wavelengdis. This approach produces a matrix of data which can be reduced to a vector of RI values at different wavelengdis.

The SPR excitation condition is extremely sensitive to die changes in die refractive index of die dielectric layers (e.g., substrate and sample) surrounding die SPR-supporting conductor sensing layer. The refractive index of a dielectric sample (in a sensor wim a given substrate) can be detected by monitoring the SPR condition. Any shift in resonance curves (66^ or δλ_spr) indicates a change in refractive index at die conductor/ sample interface.

SPR measures die complex refractive index of me sample in contact wi die sensing area of d e lightpipe. This complex refractive index includes bodi d e real and imaginary refractive index components. The real component is inversely proportional to die speed at which light propagates dirough d e sample, and is generally considered die "true" refractive index of die sample. The imaginary component is related to the sample's absorbance or attenuation of light. SPR sensors can thus be used to measure die absorbance of a sample as well as its index of refraction.

The SPR sensors of diis invention employ a planar lightpipe configuration for zero or first order sensor applications. As used herein d e term lightpipe relates to a diree dimensional strucmre diat confines optical energy and allows it to be conducted from one point to anodier point widi mmimal loss by total internal reflection (TIR). The dimensions of a lightpipe are large in comparison to die wavelength of light to be confined and conducted therein, so d at a lightpipe allows a continuous range of directions of light propagation widiin its boundaries. In addition, due to its relatively large dimensions die propagation of light witiiin a lightpipe can be modeled using ray theory ratiier dian mode tiieory. In contrast to a lightpipe, a waveguide is a tiiree dimensional strucmre that confmes and conducts light having dimensions comparable to the wavelengdi of light it confines and conducts. Light in a waveguide retains modal properties; it may be single-mode or multi-mode, but die allowed wave vector values are not continuous as in a lightpipe. The propagation of light in a waveguide, except when diere are a very large number of modes, i.e., when die waveguide is almost a lightpipe, can be modeled using mode tiieory but not ray theory.

A first order SPR planar lightpipe sensor configuration of tiiis invention is illustrated in d e side view of Figure 3A and the perspective view of the lightpipe sensor of Figure 3B. The sensor comprises a planar lightpipe (20) having top (21) and bottom (22) planar surfaces, sides (23) and input (26) and output (27) end faces. A lightpipe is fabricated from a dielectric substrate material that is substantially transparent to input light. For example, various types of glass and crystals, including sapphire, and various types of plastics and polymers can be employed as d e lightpipe materials. The planar lightpipe can be a uniform slab of substrate with uniform length (L), widdi (W) and thickness(f) (as shown in Figure 3B), where the length is defined as the dimension along which light traverses the lightpipe. Alternatively, the longimdinal input and output ends can be beveled. The lightpipe is preferably substantially longer dian it is tiiick (L> >t). In the configuration of Figure 3 A, light is end-coupled into and out of die lightpipe. The longimdinal end faces (26 and 27) are polished at about 90° (substantially perpendicular to the planar top and bottom surfaces). The SPR sensing area (30) of the illustrated sensor is fabricated on one planar surface of the lightpipe, which is designated as die top surface in all configurations described herein. The sensing area of length (1) and width (w) is shown as placed centrally on d e lightpipe surface in Figure 3B. A sample cell (31) allows the sensing area to interface with a dielectric sample. Collimated white light, from light source (28) is asymmetrically coupled into the input end of the lightpipe, for example by focusing the light, through a cylindrical lens 32 at the input face at a range of angles (θ) filling angles on only the lower side of die optical axis of die lightpipe (25). The input light must comprise TM polarized light, but die TE component of the light need not be removed. To reduce noise levels at the detector, it is preferred in all SPR configurations herein that the TE component of the light be substantially removed prior to light detection.

The sensing area comprises a SPR-conducting layer, and optionally comprises an adherence layer, a reactive layer, a dynamic range controlling layer or combinations thereof.

The term "lower side" is used in reference to the top and bottom lightpipe surfaces as defined above. The optical axis of d e planar lightpipe extends die length of d e lightpipe passing dirough d e center of the input and output faces. That axis defines a plane cutting dirough the center of the lightpipe parallel to its top and bottom planar surfaces. The term "asymmetric light input" is used herein when light entering die lightpipe at a range of angles asymmetric with respect to that plane and consists of a range of angles either below that plane or above tiiat plane, but not botii. In contrast, as discussed below, "symmetrical light input" refers to light input that is symmetrical with respect to die axis plane with a range of angles both above and below die plane.

When die planar lightpipe is asymmetrically illuminated as illustrated in Figure 3A, the output consists of discrete angular bands (b) (e.g., b = 0-9 in Figure 3A) (Smitii, W.J. (1992), Modern Optical Engineering. The Design of Optical Systems. 2nd ed.,

(McGraw Hill), p. 192). As the propagation angle (θ_b) increases down from the axis (corresponding to increasing band numbers in Figure 3 A), the direction of d e output alternates in bands above or below the lightpipe axis. This is caused by d e alternating number of reflections inside d e lightpipe, with an odd number of reflections causing die rays to be directed down, and wid an even number, directed up.

The light exiting the lightpipe of Figure 3A can be imaged on an imaging screen

(33) as shown. Other light detection schemes can be employed. For example, each angular band can be individually focused into a spectrograph to obtain die reflection spectrum as a function of wavelength. Each discrete angular output band contains a different white light SPR reflection spectrum that represents the average over a small range of angles, which is approximately equal to t L radians.

Figure 3C illustrates the reflection spectra with SPR feature for output bands 5, 7, and 9 from the SPR configuration of Figure 3 A obtained using a water sample. These spectra were obtained using a spherical lens to image each band into die input of a fiber optic spectrograph detector. The number of angular bands that can be collected and detected is dependent upon the specific optical system design.

In order to optimize the geometry of a lightpipe sensor and optimize die location of sensors on d e lightpipe for a particular application, it is necessary to know the angular range of each band and the locations at which the band reflects off of die top surface

(where the sensing layer will be positioned.) The angular extremes of each band can be calculated by "unfolding" die lightpipe multiple times as described in Smitii, W.J. (1992), Modern Optical Engineering. The Desipn of Optical Systems. 2nd ed., (McGraw Hill), p. 192. When this is done, die light rays appear to straighten out, as shown in Figure 4, and it can be seen diat the different bands hit different locations along the length of the top surface, and tiiat the extreme rays for each band are defined by die lines from the source spot to d e top and bottom of the aperture formed by die virtual lightpipe end. For a point source located at die center of the input face, the steepest angle for a band "b" is given by:

where t is the lightpipe thickness, L is the lightpipe length, and b is a positive integer which refers to the band number determined by die number of internal reflections. The shallowest angle in a band corresponds to die steepest angle in the previous band, (θmiii_b = θma _b.₁). Therefore, die range of angles (Δθ = θmax-θmin) subtending each band scales witii t L. For input white light, as Δθ is made smaller, by increasing L with respect to t, each band becomes more like a collimated white light beam, which has me effect of reducing die width of die resonant dip.

The same unfolded diagram can be used to find where each band interacts widi die sensor area on the top surface. Referring again to Figure 4, and assuming a point source located at d e center of die input face, the location where a ray passes through a virtual surface corresponds to die location of a reflection off the top or bottom of the lightpipe. The location where the b^Λ band makes its r"¹ reflection off die top surface is calculated using the relationships for A_{b r} and B_{b r} in equations 5 and 6. At die point of reflection, the diverging band spreads across die top surface from A_{b r} to B_{b r}, where A_{b r} is the horizontal distance from die input end to die reflection location for the steepest angle and B_{b r} is die distance for die shallowest angle.

_ (2r- 1.5)1 ₍₅₎

_β _ (2r-1.5)L (6) ^6,r b+0.5

If A_{b r}< x<B_{b f} for any surface location x, dien that location is illuminated by the r^tb reflection of the b^ώ band. The location and size of die top surface reflections for the first 14 bands of a light pipe are shown as horizontal lines in Figure 5. The horizontal axis indicates percent of die lightpipe length, and die vertical axis is the band number "b. " The illuminated regions extend for the length of d e horizontal line segments shown, whereas die widdi of d e spot (w in Figure 3B) depends on die widdi of the beam used in die experimental setup. For example, the incident light can be focused to a point on die input face or a line diat extends across the input face of the lightpipe. In the equations above, the illumination locations for a given band are independent of d e lightpipe diickness, and die lengtii can be factored out to leave a fraction of die total length. Thus, the reflection pattern in Figure 5 is independent of tiie light pipe dimensions.

The pattern in Figure 5 can be used to optimize the sensing area (conductor or metal layer) locations and size for a specific sensing application. Some first order sensing applications require obtaining resonances from a maximum number of bands. In such a case, a band witii partial or multiple reflections off the sensing film would be acceptable as long as a resonance was observed. For example, in measuring die dispersive RI, it is preferable to analyze as many different bands as possible each having a different resonance at different wavelengths to allow determination of RI at many wavelengths. In another example, many different components of a given sample can be simultaneously assessed using differential functionalization of die SPR sensing areas (i.e., providing different reactive layers) on a lightpipe sensor surface. In this case, each exiting angular band would carry information about a different component (analyte) in the sample. In yet another example, having the ability to analyze multiple bands sensing die same sample or analyte provides several benefits. It allows an internal confirmation or check of a given measurement, thus avoiding spurious readings d at might occur on a single channel. Further, since the higher the band angle, die greater its sensitivity and die less its dynamic range, the availability of multiple bands allows a selection of die highest sensitivity channel for a given sample.

Omer applications might require high sensitivity of Δ λ_SPR/ΔRI in a limited number of bands. These applications would benefit by designing die sensor to operate at the highest angle possible, witii only one or two reflections per band. For example, a bioassay for an analyte present at a very low concentration, where the dynamic range was known, preferably would be done employing a single band widi as high a sensitivity as possible. A large number of high angle bands are made available, for example, by placement of the SPR-sensing area within die first about 20% of die lengtii of the lightpipe. In such a configuration, low angle bands will not hit the sensing area placed in this region. In addition, sensing area placements can be selected for interaction of a particular band with a particular sensing area. For example, a narrow sensing area located near the 50% L mark would be optimal for band 13, but adjacent bands would miss the sensing area.

In addition, die pattern in Figure 5 can be used to position multiple sensing areas along die length of the lightpipe such that different bands of light will interact with different sensors. The individual sensing area can, for example, have different selectivity for interaction with species in the sample.

In order to maximize the number of useful bands, die optimization is an iterative process based on observing the pattern in Figure 5, and positioning the sensing area so that it reflects as many bands as possible. In general, d e region near die beginning of the lightpipe contains a reflection from most bands and has wide spacing between reflections.

The positioning of sensing areas in this region requires looser fabrication tolerances.

Experience has shown that bands must have at least one half of a reflection off the sensor to provide a useful resonance, and tiiat three or more reflections off d e sensing area tend to broaden die resonance and flatten its bottom, complicating die determination of λ_SPR.

In the case of designing for the highest sensitivity, die first step is to calculate the total internal reflection (TIR) angle, θ_cπucal where:

n cladding (6)

^critical = ∞∞"^{1 n} substrate

where n,,,,,..^ is die index of refraction of die lightpipe substrate and ti_di_ng is die index of die medium in contact with the substrate surface at the reflection location, e.g., air or sample. The bands closest to die critical angle are then used. The wide spacing between reflections at die beginning of die lightpipe lends itself to selecting a sensing area where d e first illuminated spot for each band is completely covered (in Figure 5) with the proposed sensing area to provide a single, complete reflection.

To compare theory with experiment, ray tracing was used to count the number of times each ray, as illustrated in Figures 4 and 5, was reflected from the sensing area of an experimental sensor of Figure 3 A (see Example 1). The graph in Figure 6 shows me calculated number of reflections that a ray of given angle (by Band No., as illustrated in Figure 3 A) makes off of a sensing area located on the planar lightpipe surface as in Figure 3 A. In Figure 6, rays contained in odd numbered bands are indicated widi tiiick lines, while rays contained in even numbered bands are indicated with thin lines. The calculations based on ray theory in Figures 5 and 6 predict tiiat band 3 will only have a partial reflection off die sensing area, almost missing d e sensing area entirely, and tiiat bands 5, 7 and 9 will respectively undergo 1, 1, and 2 complete reflections off the metal sensing area. The graph of Figure 6 also indicates the range of angles in each band. A double reflection squares the reflection coefficient, causing the resonance to broaden out, and giving a broader dip in tiie spectral intensity. A resonance on band 3 could not be experimentally measured in the test sensor. As illustrated in the spectra of Figure 3B, the resonance spectrum for band 9 is significantly broadened compared to die spectra for bands 5 and 7. The observed spectra confirm the model's prediction of die number of times each band hits d e metal surface. However, the model is sensitive to the position at which light is focused on d e input face. The model used assumes that the rays originate at the center of the input face of the lightpipe. It was found experimentally that when light is focused elsewhere on the input face, the angular content of each band shifts. For example, if the spot of focus moves from the center to the bottom on the input face, the angles in each band will shift over one half of a complete band worth of angles. A slit or aperture on or at the input face can constrain die input light to one location. These factors should be taken into consideration when using the model to position a sensor for a given application.

If light is introduced into the lightpipe of Figure 3A symmetrically at angles above and below the axis of the lightpipe, the light focused on die lightpipe input face from above die lightpipe axis will also give a pattern of angular bands on exiting me lightpipe. This lightpipe configuration is illustrated in Figure 7 where a cone of angles of light is focused at input face 26. The cone is shown focused at die center of the input end face. The input optics of Figure 7 include two cylindrical lens 34 and 35. Lens 35 collimates light across the widdi of d e substrate and lens 34 focuses a range of angles relative to the substrate thickness on die input face. The configuration includes an aperture (29) to minimize input of undesired angles of light. The output angular bands are indicated and labeled by band number and T or B (e.g., IT, IB, etc.) to indicate diat a band originated from input angles above (T) or below (B) me lightpipe axis. The banded output pattern from light of angles above the axis (IT, 2T, 3T, etc.) will be the complementary pattern of diat from light focused at angles from below die axis (IB, 2B, 3B, etc.). If die lightpipe planar surfaces are perfectly flat and, if the exit end is polished perfectly flat, the complimentary pattern will fill the gaps between the light bands in die pattern generated by light focused from below the axis when viewed near die output face of die lightpipe.

If die lightpipe strucmre is not perfectly flat, the input end is not perfectly flat and/or die input light is not perfectly focused to a point, mere will be some overlap of exiting bands (as indicated in the Figure 7) in the symmetric light input configuration. The exit optics and detection scheme used must take into account potential overlap of d e angular output bands. Each angular band can be individually focused into a fiber optic pickup for individual sequential analysis or two or more angular bands can be analyzed simultaneously witii a multiple channel detector. The banded output pattern seen in side view in Figure 7 is die result of die range of illumination angles provided by lens 34 and does not rely on whether die light is in fact collimated by lens 35.

Light input is shown in Figure 3A as completely asymmetric (all angles below the axis) and in Figure 7 as completely symmetric. Light can, of course, be introduced in a manner between these extremes. There is no requirement that the same range of angles above and below be introduced. The use of die asymmeϋric light input scheme illustrated in Figure 3 A avoids potential overlap of angular bands, simplifying output detection, and is in this sense preferred. However, if angles on both sides of die axis are filled, there are more output bands that are closer together, thus providing more choices for optimization.

The discussion and calculations illustrated above regarding optimized placement of me sensing area in the lightpipe assume asymmetric light input below the optical axis of the lightpipe and placement of the sensing area, by definition, on die top planar surface of the lightpipe. As noted above, the lightpipe sensor can be configured widi symmetric light input (Figure 7). The lightpipe sensor can be also be configured with asymmetric light from the top side of die lightpipe, i.e. light input filling angles above die lightpipe axis, witii the sensing area still located on the top surface of the lightpipe. Optimization of placement of the sensing area in these cases is done using procedures and calculations analogous to those discussed above.

The lightpipe of Figure 3A is shown with both input and output ends perpendicular to the top and bottom surfaces. Both the input and output ends of die lightpipe can be beveled, i.e., angled with respect to the top and bottom surface. Figure 3 A shows asymmetric light input of a wedge of angles from below the axis. (Figure 7 below shows input of a cone of light on a perpendicular input face.) Light can be coupled into d e lightpipe at a range of angles by focusing a cone or wedge of light (as in Figures 3A and 7) at d e beveled or angled input face. Those of ordinary skill in die art will appreciate that an input configuration equivalent to the input configuration of Figure 3A can be achieved by directing the wedge of light along die optic axis of die lightpipe and focusing die light at a beveled edge, i.e., die equivalent of off axis input can be achieved widi on axis input to a beveled end. The input face can be beveled so tiiat it makes eitiier an acute or an obtuse angel with the bottom planar surface of d e lightpipe. The choice of bevel angle in on-axis light input allows a selection of a range of input angles both above, or below (or both above and below) d e axis.

One particularly useful means for collecting the output light of the lightpipe is illustrated in Figure 8. In diis side view, an SPR sensor of Figure 3 A incorporates a telecentric system wherein cylindrical lens 40 is placed at about one focal length from the output end (27) of the lightpipe. In this configuration lens 40 is a telecentric lens and die lightpipe output end (27) is a telecentric stop. The lens collects light exiting the lightpipe and focuses it into the detection optics (Smidi, W.J. (1990) Modern Optical Engineering. The Design nf Optical Systems. 2nd ed., McGraw Hill, N.Y.).

A telecentric system comprises a telecentric lens which may itself have several components. A telecentric lens is a lens in which an aperture stop (the telecentric stop) is located at die front focus and results in die chief rays of light passing through the lens being substantially parallel to the optic axis of the lens in image space. In the illustrated telecentric system, a cylindrical lens simultaneously redirects and focuses output bands onto a plane perpendicular to die lens axis and makes the bands substantially parallel to d e optical axis of the lens. A spherical lens placed in an analogous telecentric strucmre (i.e., a spherical telecentric lens) simultaneously redirects and focuses die light from die bands (as would be seen from both the side and top views of die substrate, not shown). A telecentric lens images output bands into individual spots or lines, so that a detection device, e.g., fiber optic pickup, photodetector or entrance slit for a spectrograph, can be placed at die image location (about 2 focal lengtiis from the telecentric lens, 2f) to capmre the signal with minimal effects from defocus. The telecentric lens is positioned one focal length from the output end of the lightpipe sensor so that the output face of die sensor acts as a telecentric stop (41). Alternatively, albeit with significant light loss, a separate output aperture can be provided to reduce die angular width of each band. Since each band of light passes dirough an aperture at die focus of the lens, die padi of each output band is straightened out to be substantially parallel to the optical axis of the lens.

The individual bands exiting d e lightpipe can be considered to originate from locations along a virtual arc (42), so that each band will come to a focus some distance behind (2f) die telecentric lens. When d e telecentric lens has a focal length approximately equal to the distance between the virtual source origin to tiie end of die lightpipe, die size of die output focus and spot will be comparable to the size of die incident spot. This choice of focal lengtii for die lens is preferred for efficient light collection by the detection optics. The telecentric lens is preferably chosen so diat die field curvature of die lens is matched to the profile of the virtual arc of sources (Smith, W.J. (1990) supra and Optics Guide 5, Melles Griot Catalog (1990)). When the condition is met, all the bands will focus at a common plane perpendicular to die optical axis as shown in Figure 8. This choice of lens is particularly useful and preferred for capturing multiple bands of light at d e same time. It is also preferred when capturing multiple bands tiiat the lens is chosen and positioned so tiiat the separation between the center of each band is constant. Then the focus of the light into the detector optics can be adjusted by moving the strucmre (sensor and lens) along d e axis without d e need for transverse adjustment.

In Figure 8, the optical axis of the telecentric lens is aligned widi the optic axis of the lightpipe. The lens can be employed in odier output optical geometries for capmre of angular bands. For example, the telecentric lens axis can be aligned witii the center ray of any angular output band to capmre adjacent bands. In this case, the telecentric lens will redirect and focus d e bands passing dirough it to make them substantially parallel to the optical axis of the lens (i.e., parallel to the direction of die output band to which the lens is aligned).

The output optics of die lightpipe of Figures 3A, 7 and 8 preferably include a means for correcting for spherical and chromatic aberrations so that a good image of the source corresponding to each angular band can be formed. This type of correction can be accomplished for example using a field flattener.

Since it is possible to calculate the precise range of angles in each band of light (exiting d e lightpipe or propagating through the lightpipe), the lightpipe sensor of Figure 3A in which the sensing area consists of an SPR- supporting conductor layer, specifically an SPR-supporting metal layer, of known tiiickness can be calibrated using a model matching technique. By measuring d e length and tiiickness of the lightpipe with a micrometer, the angles of each band can be calculated wid high precision. If the thickness of the metal sensing layer is accurately known, a Fresnel reflection model can be used to find die RI which would excite SPR at the measured resonance angle. If die metal tiiickness measurement is not accurate, metal thickness in the model can be corrected by using a liquid standard whose RI is measured on an Abbe refractometer or by other appropriate methods. Metal tiiickness can be adjusted in die model until the modeled resonance at the wavelength at which the standard is measured matches die experimental resonance of d e liquid standard.

An alternative first-order planar lightpipe sensor configuration is illustrated in Figure 9A. As in the sensor of Figure 3A, the planar lightpipe (43) has a top (44) and bottom (45) surface, sides (46) and input (47) and output (48)ends. In die illustrated configuration, d e input and output ends are symmetrically beveled at a selected angle α to die bottom surface by polishing. In a specific embodiment the input and output bevels are each at a 45° angle to the bottom surface of the lightpipe. Again in the illustrated configuration, the upper planar face of the sensor carries the sensing area (30) in which SPR can be excited and which contacts d e dielectric sample in a sample cell (31). The sensing area comprises an SPR-conducting layer and optionally comprises an adherence layer, a reactive layer, or a dynamic range-controlling layer. Light from a white light source 49, and which is optionally passed through a TM polarizer, enters and exits die lightpipe at its input and output ends, respectively, dirough the bottom planar face of the lightpipe. In this configuration the light source and detector system (50) are positioned away from the sample so that the chemistry and optics can be separated. Collimated white light is focused to a point by a spherical focusing lens (51) (or a line using a cylindrical lens) at die bottom surface of the lightpipe near its input end with a range of angles θ with respect to the bottom of the light pipe. The light is focused such tiiat it reflects off the bevel of the input face while stray light is directed away from the detector. The bevel is illustrated as about 45 ° but can be any angle d at will allow light coupling into d e lightpipe. The bevel can be metalized to act as a mirror or it can allow light to couple into the lightpipe via total internal reflection (TIR ) off its surface. The light entering the lightpipe in such a configuration need not be and is preferably not perpendicular to the bottom surface of the lightpipe. The range of input angles of light is limited by TIR inside the lightpipe.

The output bevel (also shown as 45° with respect to the bottom surface in the figure) directs d e light down towards die input side of d e sensor into a detector system 50. A diffraction grating (53) index matched onto die output bevel reflects light out of die lightpipe dirough the bottom surface. The grating is oriented to disperse each band out of d e page and a portion of the dispersed bands are collimated witii cylindrical lens (55) and imaged with cylindrical lens (56) onto an output detector plane for detection, for example by an imaging detector (57) to produce an image of reflected light intensity vs. angle and wavelengdi. Alternatively, a diffraction grating can be formed directly on the beveled output end so d at no index matching is necessary. Further, die diffraction grating can be replaced witii a mirrored surface, e.g., the a mirror can be deposited on die beveled end.

The inset graph of Figure 9B illusurates an idealized contour plot of reflection coefficient of the output light as a function of both wavelength and angle. Cylindrical lens (55) is positioned to affect only those bands exiting the lightpipe that have undergone an odd number of reflections. The complementary pattern produced by die even bands is missed by lens (55). The output spectra can be analyzed to determine λs_PR for each band. The measured center angles of each band are plotted against λ_spr in Figure 10 and compared to values calculated using a model discussed in Example 3. Figure 10 shows tiiat Δλ_spr/ΔRI measured using die sensor closely matched values predicted by tiieory.

A specific embodiment of the planar lightpipe single-angle SPR sensor of this invention is illustrated in Figure 11. Figure 11 gives a side view of die lightpipe widi optical component for light input and detection. In this zero-order sensor, d e planar lightpipe 61 of length, L, and tiiickness, t, where L is the length of the longer of the top or bottom surface, is beveled at a selected angle α at botii its input (62) and output (63) end. Both beveled surfaces can be mirrored (64). The lightpipe has a top planar surface (66) carrying an SPR sensing area (68) comprising an SPR supporting-conducting layer (69). The sensing area optionally comprises an adherence layer, between the lightpipe surface and die conducting layer and optionally comprises chemically or biochemically selective overlayers, reactive layers, which can provide for selective adherence of analytes from the sample. In addition, the sensing area optionally comprises dynamic range - controlling layers, and passivation or protective layers. The sensing area of length (1) is symmetrically positioned along die lengtii (L) of tiie lightpipe. The sensing area can extend over the entire width of the lightpipe or over a portion of that widdi. In the sensing configuration, the sensing area is in contact with a dielectric liquid or gas sample, for example in a static cell (70) or flow cell configuration (not shown).

A collimated beam of white light comprising TM polarized light (preferably TM polarized light) is coupled into die lightpipe dirough die bottom planar surface and reflected from the internal surface of the input bevel. As illustrated, light propagates dirough die lightpipe by TIR to hit die sensing area at a single fixed angle, θ, which is 2α, and excite SPR at the sensing area. Since light cannot in a practical sense be perfectly collimated, d e light hitting the sensing area comprises a small range of angles dependent on the quality of input collimation. The output bevel is symmetrical to the input bevel and couples light downward out of die lightpipe. This lightpipe structure was designed to be symmetrical so tiiat metal sensing layer is located in d e center of die top surface. The bevel angles (α) were chosen so tiiat the incident angle of illumination on die active sensing layer (θ = 2α) would be high enough to ensure good sensitivity. The angle that ensures good sensitivity depends on d e substrate used and the particular application. For glass and low concentration aqueous samples, 22°-25° or 75°-78° from the normal is a good balance between high sensitivity and die resonance occurring past die wavelength range of a silicon detector.

The sensor of Figure 11 comprises a non-monochromatic source, preferably a broad band or white light source. Light exiting die source (73) is collimated using collimator (71) and is optionally TM polarized using polarizer (72). The light may be collimated as a narrow beam that lands on a point on the center of the face of the input bevel. Alternatively, the collimated beam of light can be expanded in one direction using anamorphic optics so that it forms a collimated line across the width of the input bevel face. Collimated light enters die lightpipe, reflects down die lightpipe interacting with the sensing area, exits the lightpipe by the output bevel and is collected by output signal collector (74) (a collimator lens used backwards to image the collimated light into the fiber optic pick up of a spectrograph). The collected output then passes to a suitable detector (75) to measure reflected light intensity as a function of wavelength, e.g., a spectrograph. Assuming a symmetrical lightpipe profile to position the sensing area in the middle along die length of die lightpipe, die lengtii (L) of the lightpipe is chosen using equation:

L = tcot(α) +ttan(2α) +2(2n+l)ttan(2α) (?)

where n is an integer related to die number of reflections inside die lightpipe and t is die tiiickness of tiie lightpipe. The length (1) of the metal sensing layer is preferably selected so tiiat light hits the sensing layer only once. The preferred maximum widdi l,_^ of die sensing area to meet diis condition depends upon t and α and is given by:

1 = 2t tan(2α) (8)

Exemplary reflection spectra obtained wid d e SPR sensor of Figure 11 are provided in Figure 12(a)- 12(d). See Example 3 for details of the measurements with die sensor of Figure 11.

In alternative embodiments, diffraction gratings fabricated on die planar lightpipe can be employed at die input and output ends of die SPR lightpipe sensors to couple incident monochromatic light into or reflected monochromatic light out of me lightpipe. A diffraction grating can be introduced onto die bottom or top surface of die lightpipe near eitiier the input or ouφut end, or at both ends. The grating is created in or on die substrate material of the lightpipe by conventional methods, for example lithography and etching techniques standard in the semiconductor industry. Light is coupled into die lightpipe at desired transmission angles by focusing light at the appropriate incidence angles onto die input grating.

The SPR sensor planar lightpipes of diis invention comprise a sensing area adhered to d e external top surface of d e lightpipe. Detection of a sample or a given species in a sample by the lightpipe SPR sensor is made, in part, by contacting the sensing area of the lightpipe with the sample. The sensing area is prepared by adherence of an SPR- supporting conductive layer to a selected area on an external longimdinal surface of die lightpipe. The position and length of the sensing area is selected to optimize die sensor for a given application.

The lightpipes of d e SPR sensors of this invention are fabricated from a material that is transparent or semi-transparent to the range of wavelengths of light to be employed in a given application. Useful substrates include glasses, crystals, plastics and polymers. To insure TIR in the lightpipe, e.g., along the length of the lightpipe, the lightpipe is optionally provided witii a cladding layer having an index of refraction different from tiiat of die lightpipe substrate. The cladding is provided over die entire lightpipe (except for sensing area) or over selected portions of die lightpipe. Those of ordinary skill in the art know and understand how to select and can readily select a lightpipe substrate appropriate for a given application. Those of ordinary skill in the art also know how to select and can readily select an appropriate cladding layer for a given application and substrate material.

The sensing area comprises one or more layers which together support SPR. The sensing area comprises an SPR- supporting conductive layer. This layer may be a conductor, e.g., a metal layer that supports SPR or a semiconductor layer diat supports SPR. Semiconductors useful in die conductive layer include silicon and germanium. Alternatively, conductive polymers can be used in die conductive layer.

The conductive layer can be a "SPR-supporting metal layer" which is herein means a highly-reflective metal that supports SPR at the metal/sample interface and has a permittivity constant wherein the real part of die permittivity is negative and its magnimde is greater than d e magnitude of die imaginary part. For wavelengdis in die visible and near-infrared (i.e., 400 nm-1000 nm), both silver and gold satisfy diis criterion. The SPR supporting metal can also be a mixmre of one or more metals or be composed of sequential layers of different metals. If die wavelengdi range utilized extends into the infrared, otiier metals, such as aluminum, copper and tantalum, may also be used. Preferably the SPR-supporting conductive layer, e.g., die metal layer, is adhered to d e lightpipe surface to a thickness which will optimize the measured resonance curve, i.e., to a thickness which makes the SPR resonance spectrum both deep and sharp, between about 400 A to 700 A tiiick. When the SPR-supporting metal layer is made of silver, the layer thickness preferably is between about 500 A to 550 A thick. Layers of silver thinner than about 400 A result in substantially shallow and broadened resonances, and layers tiiicker than about 600 A will result in significant diminishment or disappearance of the resonance feature. The range of thicknesses for gold SPR-supporting layers are also 400-700 A, preferably 500-600 A. Gold is preferred because of its inertness and resistance to oxidation. SPR-supporting metal layers can be prepared wid sequential layers of different metals, for example, a base layer of silver combined witii an upper layer of the gold for a total double layer tiiickness of between about 400 A to about 700 A. One of ordinary skill in the art can readily determine die appropriate thickness of the SPR supporting metal layer for a given lightpipe sensor application by varying die metal layer thickness to optimize the resonance curve.

SPR-supporting conductive layers are adhered to d e lightpipe surface by methods known in the art. An SPR supporting metal layer can be adhered by standard procedures, including vacuum deposition, electron beam deposition, sputtering, chemical vapor deposition and the like. Layer thickness is controlled by well-known methods, for example employing a quartz crystal oscillator or other suitable thickness monitor. U.S. Patents 4,997,278, 5,064,619, 5,351,127, and 5,485,277, for example, disclose, reference or summarize methods for adherence of an SPR-supporting metal layer.

Prior to adherence or deposition of die conducting layer a base or adherence layer is optionally applied to die substrate (here, lightpipe) surface. The adherence layer is typically a metal layer, such as chromium, nickel, platinum or titanium, less dian about 50 A thick, and more preferably about 20 A thick.

The sensing area optionally contains one or more additional layers adhered to die

SPR supporting conductive layer to yield a change in die effective refractive indices detectable by die sensor. Such additional layers can include a dynamic range-controlling layer, a reactive layer, a protective overlayer or any combination mereof. A variety of techniques are known and available to tiiose in the art to provide dynamic range- controlling layers, reactive layers and protective layers in an SPR sensing area.

A "dynamic range-controlling layer" is a layer adhered to die SPR supporting conductive layer to alter the dynamic range of die SPR sensor. This layer has an index of refraction different (eidier higher or lower) than that of tiie SPR-supporting layer. For example, adherence of a layer of higher refractive index to die index of die substrate will extend die dynamic range of the sensor to include lower RI values. For example, U.S. Patent 5,327,225, describes the use of an overlayer of relatively high refractive index material, specifically SiO, on a fiber SPR sensor with a silver SPR-supporting layer to shift the dynamic range of the sensor to a lower RI value.

A "reactive layer" is an optional layer in the sensing area which interacts with a sample or an analyte species in the sample such that the effective refractive index detected by the sensor is altered. The addition of die reactive layer permits die manufacmre of an SPR sensor which is more sensitive or selective for a sample (or analyte in a sample). Suitable reactive layers include those used in biological sensors, e.g., an antigen, antibody, nucleic acid or protein bound to die SPR supporting metal layer. This type of reactive layer will selectively bind a species in die sample, for example, a cognate antibody or antigen or complementary nucleic acid in die sample, increasing d e thickness of the reactive layer and causing a shift in the effective refractive index measured by die sensor. Most generally, suitable reactive layers are altered in some way by contact with the sample so that die effective refractive index as measured by die sensor is changed. Reactive layers also include sol-gel films and polymer coatings. Reactive layers can be adhered to die SPR-supporting conductive layer or to an overlayer on the conducting layer. The reactive layer should interface with the sample solution.

U.S. patent 5,055,265 and 5,478,755 relate to SPR sensor configurations utilizing so-called "long-range SPR" (LRSPR). LRSPR differs from traditional SPR in die use of a distinct layering in die SPR sensing area. LRSPR employs a thinner conducting layer (100-200 A) than in traditional SPR (500-600 A). An LRSPR sensing area is fabricated by first depositing a diin dielectric layer on die transparent substrate after which the diin conducting layer is deposited. The metal layer can directly contact die dielectric sample, or a reactive layer can be laid down upon the conductive layer. LRSPR in general provides increased sensitivity. The sensor configurations of this invention can be employed for LRSPR by appropriate adjustment of die layers of the sensing areas.

Printz, M. et al. (1993) J. Modern Optics 4Ω (11):2095-2104 and Bussjager, R. and Macloud, H. (1995) J. Modern Optics 42(7): 1355-1360 have described a variation of SPR diat is designated "inverted SPR". These references are incorporated by reference herein for their description of "inverted SPR." This method differs from traditional SPR in that the SPR-sensing layer comprises a thicker layer (about 100 A) of a metal, like chromium, which is usually used in an adherence layer, witii a thinner layer of gold or silver (about 400 A) on top (i.e., for contact with dielectric samples). The SPR signal has an inverted feature in it and the wavelength of the resonance for a given sample is shifted from that measured by SPR. The sensor configurations of this invention can be employed for LRSPR by appropriate adjustment of the layers of the sensing areas.

A number of methods have been described, are known and available to tiiose of ordinary skill in the art, for the formation of reactive layers with sensitivity to a variety of biological or chemical species. Formation of the reactive layer on a metal layer may require an intermediate diin layer of material to passivate the metal or protect ligands in die reactive layer from reaction with the metal; For example, U.S. patents 4,844,613, 5,327,225, 5,485,277, and 5,492,840 disclose or summarize memods for preparation of such reactive layers in SPR sensors.

An SPR sensor of this invention can be configured with one or more sensing areas. One or more active sensing areas (those capable of detecting changes in RI of a sample) and one or more reference sensing areas can be provided in an SPR sensor. Active sensing areas in an SPR sensor can be provided witii different reactive layers (e.g., can be functionalized for interaction with different biological or chemical species or functionalized differently for interaction with the same biological or chemical species), different over- or underlayers, different dynamic range-controlling layers and/or combinations thereof.

A sensing area on a planar lightpipe of this invention can, for example, be subdivided into lateral regions across its width to provide separate sensing channels, including reference channels and sensing channels with different analyte selectivities. Differential sensitivity can be provided by use of different reactive layers. One of die lateral regions of the lightpipe can function as a reference for other activated and functionalized sensor channels. If die reference region is not functionalized or activated (i.e., no reactive layer provided) it will serve to track temperamre changes, variations in the light source and signal due to nonspecific adsorption of d e analyte. Alternatively, the reference area can be coated witii a thick layer of a reference material (a dielectric) so tiiat it does not react to d e sample and will serve to track temperamre changes or variation in the light source.

A sensing area or a portion of sensing area can also comprise an overlayer that protects or insulates the SPR-supporting layer from changes in the RI of the sample. For example, a reference sensing area can be made by providing a sufficiently thick overlayer of a dielectric material, such as a cured epoxy on die SPR-supporting layer. The reference sensing area then senses, and be used to correct for, temperamre variations, light source variations and related instrumental variations.

In practice, the refractive index along the sensing interface in an SPR sensor is temperamre dependent, as are die characteristics of d e light source and the detector in d e sensor system. Temperamre fluctuations lead to variation in die SPR excitation condition and cause undesirable shifts in SPR wavelength. Thus, temperature compensation of the SPR sensor can significantly improve the accuracy of sensor measurements. One way to deal with temperature variation is provide a means for keeping the sensor at a constant known temperature (e.g., temperature control). This may not be practical in certain sensor applications. Alternatively, a sensor can be temperamre compensated by developing a complex algorithm to allow correction of sensors measurements as a function of temperature variations. A third method for accounting for temperamre variations to improve sensor accuracy is to incorporate a reference SPR signal as the compensation mechamsm. SPR sensor configurations employing the planar lightpipe of this invention are readily adapted to include multiple sensor channels, one of which can be employed as a reference for temperamre compensation.

Figure 14A is a side view of a planar lightpipe employed in SPR sensors of this invention. Figure 14B is a top view of the sensor of Figure 14A showing that the sensing area on the top surface of the planar lightpipe is divided into two parts along the width of die top surface, which constitute two different sensor channels (81 and 82). Botii channels are fabricated on die same sensor substrate, for example on a standard microscope slide, and have an identical SPR- supporting conductor layer and optional adherence layer. The length and positioning of the sensing areas are selected as discussed above. Only one of die channels is active for sensing particular species in d e sample solution. The sensing areas differ in the functionalization of the conducting surface for interaction with specific species in the sample. For example, only one of the sensing areas (81) contains a reactive layer, as described above, which interacts witii a specific species in solution causing a shift in SPR resonance. The other sensing area, tiie reference sensor channel (82), is inactive to such interactions and SPR on this sensor channel is a function bo of the effective RI of the sample and non-specific absorption events.

The configuration of Figures 14A and 14B is similar to that of Figure 7. The optional aperture 83 has been adapted for use with two sensing areas. This configuration is illustrated witii a flow cell 85. The detection optics include a cylindrical lenses, like 86 to collimate bands and spherical lens 87 (a, b) to image bands into fiber pick ups (88 a,b). As illustrated, the detection optics can be configured to conduct individual bands into individual fiber pick ups which ultimately lead to a detector. For example, an array of fiber optic pick ups can be provided to detect a plurality of angular bands, or one (or several) fiber optic pick ups which can be adjusted to pick up any desired angular bands can be provided.

As long as die functionalized (reactive) layer in the active layer is thin (generally these layers are only a few monolayers tiiick), the temperamre dependence of the two channels is substantially d e same. In the configuration of Figure 14B the SPR channels share the same light source and detector system. Subtraction of the SPR signals (die reference SPR from the active sensor) yields the temperamre independent system response to the interactive species. This referencing mechanism also removes the effects of light source fluctuations and system losses.

In an alternative multi-channel sensor, the reference sensing area and any active sensing areas are formed wid die identical SPR-supporting conducting layer and die same adherence layer (if any). An overlayer is then applied to die reference sensing area to provide interaction widi a layer of constant RI. For example, a relatively thick layer of cured epoxy can be used to overlay die SPR-supporting layer. The reference sensing area in this case does not respond to changes in sample RI or to any specific binding of analytes that might occur on the active sensing area(s) of the sensor. The reference sensing area responds to changes in temperamre, light source and odier possible instrumental variations.

In a multi-channel lightpipe sensor, the SPR signals from each channel can be independently collected and measured. For example, SPR signals from two adjacent channels shown in Figure 14B can be collected using fiber couplers (88a and 88b) into adjacent fibers and then transmitted into the same spectrograph, one at a time, periodically through a fiber switch under time control. Switching time and spectrograph analysis in such a system can be synchronized by a computer. Alternatively, a two-channel (or multiple channel) spectrograph can be utilized. The on-site and real time temperamre information of d e whole sensor system, thus, can be collected for die reference signal and extracted from die active SPR signal by conventional signal processing methods.

The beveled-ended zero order configuration of Figure 11 is illustrated in Figure 15 as a dual channel lightpipe sensor. A planar lightpipe was divided lengtiiwise into two sensor channels (90 and 91) adjacent to each odier as shown in Figure 14. Output from the two different sensing channels, at different positions along the widdi of the lightpipe, exits the lightpipe at corresponding positions along die width of die lightpipe and is separately collected by fiber couplers (92a and 92b), passed dirough fiber switch 93 and analyzed in a detector system comprising a spectrograph (95) under computer (100) control, the details of this configuration are discussed in Example 4.

Referring back to Figure 6, there are specific reflection locations along the length of the top surface (as well as the bottom surface, but not shown) of the lightpipe unique to a given angular band and not shared by any other band. Therefore, individual angular bands can be used to momtor SPR on different locations along the length of the lightpipe surface. Different samples can thus be placed to coincide widi d e unique different band reflection locations to create a multiplexed SPR sensor with a different signal on each band detected.

Selective placement, by choice of position as a function of both length and width on die lightpipe surface, of multiple discrete sensor areas (botii active and or reference sensing areas) on the external planar lightpipe surfaces creates a matrix of individual channels tiiat can be separately measured by selective detection of a particular angular output band at a selected position along die width of die lightpipe. For example, a matrix of photodetectors can be provided to detect die individual output of angular bands along die widdi of the lightpipe. Alternatively, a matrix of fiber optic pickups can be utilized to capmre d e output signals and relay them to a multi-channel spectrograph or a switch and a single channel spectrograph.

Both the first-order and zero-order sensors of this invention can be configured as multi-channel (two or more channels) lightpipe sensors, with sensing areas positioned laterally as illustrated in Figure 14B. Output light from the different channels is separated along the widdi of the lightpipe by the placement of sensing areas. Multi-channel lightpipe sensors have two or more sensing channels, one of which is preferably a reference channel. Different sensing channels can be provided witii different reactive layers, i.e., functionalized differentiy, to interact with die same or different species in samples. Two or more of the sensor channels can be functionalized in the same manner to interact and detect die same species in a sample to provide an internal check of SPR measurements. A multi-channel lightpipe sensor can also be combined widi multiple sample flow cells or sample cells having multiple channels, such that different samples interface with different sensing channels.

The first-order sensors of diis invention which have angular output bands, as shown in Figures 4-6, can be configured as multiplexed sensors widi each angular band sensing different analytes. The first-order sensors of this invention can also be configured witii a matrix of sensing areas to allow multiplexed sensing in different channels. Conventional surface layer deposition technology combined widi conventional masking techniques can be employed to introduce a matrix or otiier pattern of sensing areas on tiie lightpipe surface.

The SPR sensors of this invention, and particularly those exemplified in Figures 3A, 7 and 9A, can be employed as zero-order sensors by using monochromatic input light (or substantially monochromatic input light) and an appropriate detector for monochromatic light, for example, a linear detector array can be used to measure intensity vs. angular position.

The SPR sensor configuration of this invention can be operated in eitiier angular modulation mode, wavelength modulation mode or a combination of botii modes of operation. The light sources employed witii the sensor can be monochromatic or more preferably are non-monochromatic. A monochromatic light source provides light of substantially one wavelength. A non-monochromatic light source is any light source that provides light of more than one wavelength, i.e., any light source that provides multiple wavelengths. Preferably, the non-monochromatic source provides a range of wavelengdis of light sufficiently broad to encompass die SPR spectrum of the sample. A black body radiation source or one or more broad spectrum light emitting diodes are, for example, suitable multi- avelength light sources. Alternatively, two or more discrete wavelengths of light, e.g., from distinct light sources, can be employed in the sensor of diis invention. A variety of monochromatic and non-monochromatic sources of incident radiation are readily available. Monochromatic sources include laser sources, e.g., diode lasers, and gas discharge sources. In addition, a monochromatic source can be generated by coupling of white light or other multiple wavelength source with a wavelength selective filter or with a monochromator. Non-monochromatic sources include combinations of two or more monochromatic sources including, one or more LED's, arc sources, black body sources, and certain gas discharge sources, e.g., neon indicator lamps. A tungsten halogen lamp, for example, is a suitable white light source. Best results are obtained when die cuπent in, and temperature of, the white light source are controlled in order to minimize any background spectral variation.

A variety of detector schemes applicable to analysis of the output light of the sensors of this invention are known and readily available to tiiose in the art, including spectrographs, fixed linear aπay detectors, CCDs (charge coupled devices), photodiode arrays, monochromators, mechanically tunable wavelength output and a single detector, electronically tunable filters (scanning etalon), dispersing prisms and wedge etalons. For example, a photodetector can be combined witii a series of bandpass filters, e.g., a filter wheel. Passage of die light exiting die lightpipe dirough a filter wheel allows selection by rotation of the wheel of a naπow bandpass of the light for wavelength-selective intensity measurement with the photodetector. Detection systems can alternatively employ a dispersing prism, linear variable interference filters or individual interference filters when only a limited number of wavelengths are of interest.

U.S. patent 5,374,563 describes SPR sensors tiiat employ phase modulation detection. Nelson, S.G. et al. (1996) "High Selectivity Surface Plasmon Resonance Sensor Based on Phase Detection" presented at die Sixtii International Conference on Chemical Sensors (July 22-24, 1996) Washington, D.C. also described SPR sensor configurations that employ phase modulation detection. One particular difference in the use of phase modulation is tiiat the input light comprises TM and TE polarized light. The SPR sensors of this invention can be modified or adapted in view of tiiese references and what is known in die art about SPR and phase modulation to employ phase modulation detection, particular those methods specifically described in die cited references. An anamorphic lens beam expander, which is a lens system that magnifies a beam of light in only one direction, can be employed in the input or output optics of the SPR sensors of this invention. These lens systems are particularly useful for input into lightpipe sensors having a plurality of sensing areas across the widdi of the lightpipe.

In general, lens and related components employed to collimate or focus light into die SPR sensor configurations or out of die sensor of this invention are preferably achromatic.

The range of RI that can be measured witii a given sensor depends upon incident angle, substrate RI, wavelengdis of illumination detection, choice of sensing metal, dynamic range controlling layer and to some extent the metal thickness. RI values above n_glass can only be detected when using a dynamic range controlling layer.

The term "substantially" has been used to modify several absolute terms herein, e.g., substantially single angle, substantially collimated, substantially parallel and substantially monochromatic. The term is used to indicate d at some deviation from the absolute is tolerated in the configurations described herein. In some cases, for example, in "substantially collimated", die term indicates tiiat it is not, in a practical sense, possible to achieve absolute, i.e., perfectly collimated light. This is appreciated and understood by those in the art. Thus, in the "single angle" configurations of this invention imperfections in collimation of input light will lead to a small range of incidence angles at die sensing area and d e configuration will only be substantially single angle.

The term "analyte" is used herein generically to refer to any chemical or biological molecule (nucleic acid, antibody, antigen, blood factor or component, etc.) that is to be detected. The devices and methods of this invention can be used for the quantitative or qualitative detection of one or more analytes in gas or liquid samples. The device and methods of this invention can be employed in d e analysis of a solid sample or of a thin film in contact with the sensing area. The sensors of this invention can be employed in a variety of applications. In general, they can be employed in any application which currently employs a prism or waveguide SPR sensor configuration. These sensors can be adapted as discussed above for use in biological sensing applications, e.g., as biosensor, or use in flow or static sample systems. They will be particularly useful in low cost applications, such as hand¬ held SPR instrumentation. Specific examples of applications include use as a detector in instrumental effluent stream, such as in HPLC methods or for the detection of coπosion of metals.

The SPR sensors of this invention are useful in industrial process control applications, such as environmental waste stream monitoring, in pharmaceutical production and in food and beverage production.

The sensors of d is invention can be employed in combination with other analytical methods including, for example, electrochemical methods. In particular, the sensors of this method can be employed in die combined electrochemical and SPR methods tiiat have been described, for example in memods described in U.S. patent 4,889,427; in Gordon, J.G and Ernst, S. (1980) Surface Science 101:499-506 and in U.S. provisional patent application 60/007,026, filed October 25, 1995 and coπesponding U.S. patent application (Attorney Docket No. 90-95) filed October 25, 1996, all of which are incorporated by reference herein for their disclosure of combined electrochemical and SPR methods.

U.S. Patent No. 5,485,277 discloses the use of SPR sensors for enhanced fluorescence measurements. The methods disclosed combine a fluorescence detector positioned widi respect to die SPR metal layer to detect fluorescence from the layer. SPR sensors of this invention can be readily adapted witii appropriate fluorescence detectors for use in such methods.

U.S. Patent No. 5,313,264 describes die use of an optical multi-analyte sensor system based on internal reflection of polarized light in combination witii detection methods based on the evanescent wave phenomenon at TIR including SPR, critical angle refractometry, TIR fluorescence, ITR phosphorescence, TIR light scattering and evanescent wave ellipsometry. The SPR sensors of this invention can be readily adapted or modified in view of die disclosures herein, in U.S. patent 5,313,264 and in view of memods, techniques and devices d at are well-known in the art, for use in combination with TIR-based detection systems, particularly those mentioned above.

EXAMPLES Example 1.

A partially optimized sensor with the sensing area located as shown by die shading in Figures 4 and 5 was fabricated on a 28.4 mm long by 0.94 mm thick float glass microscope slide. The sensor was configured as shown in Figure 3 A, with water as the sample. The water sample was retained on slide by surface tension. Individual angular bands of light exiting d e light pipe were individually focused using a 20 mm focal length achromatic lens onto a 400 μm optical fiber and measured with a fiber optic spectrograph to obtain the surface plasmon resonance spectrum of die water sample. Light from die light source was passed dirough a TM polarizer before passing into the cylindrical lens. The spectrum of the TM polarized light was normalized using the nonresonant TE polarized spectrum to remove die lamp spectrum.

The polarizer was rotated so die light in the system was TE relative to the sensing surface. The spectrum for a band was acquired. Then die polarizer was rotated 90° so die light in the sensor was TM relative to sensing layer and the water resonance was obtained. Then each wavelengdi intensity value from TM light was divided by die same wavelength intensity value from the TE light. The resonance spectra shown in Figure 3B are the normalized spectra (light intensity as a function of λ ) for bands 5, 7, and 9 from this sensor configuration and each represents the average reflected output over a small range of incidence angles. The data collected provide resonances for the same sample at a number of incidence angles. If the individual bands are calibrated, tiiese data provide dispersive RI data for die same sample at several different wavelengths

Example 2.

An experimental sensor like that of Figure 9 was fabricated with a 0.94 mm thick float glass microscope slide A sensing area formed widi a base metal layer of 302

A of silver with an overcoat of 203 A of gold deposited on die top surface of the slide.

The sensor lightpipe was 70.59 mm long and had an 11.0 mm long metal sensing area extending the width of die planar lightpipe starting 24.5 mm from d e input end (from

24.5 mm to 35.5 mm). The sensor employed a TM polarized white light source. Polarized white light was collimated as a 5 mm beam which was focusing onto the bottom planar side of the light pipe through a cylindrical lens. The illuminated spot sizes along the top surface of the substrate were approximately 5 mm wide, widi lengths according to die illumination pattem in Figure 5. Light passing through the lightpipe and reflecting off of the sensing area excited surface plasmon waves on the metal surface and was dien reflected out of die sensor by the diffraction grating index matched onto die ouφut bevel. The grating was oriented to disperse each band out of die page. The exiting bands of light were collimated with a second cylindrical lens, and the dispersed wavelengdis were imaged witii a tiiird cylindrical lens onto a white paper screen placed a the output plane to produce an image of reflected light intensity versus angle and wavelengdi. The plot in Figure 9B illustrates an idealized contour plot of the TM reflection coefficient as a function of both wavelength and angle. Note tiiat the output of the illustrated sensor is modulated by die odd numbered angular bands of light. The complementary pattem produced by die even bands missed d e output optics and therefore were not captured or collimated.

The angular response of the system was calibrated by measuring the dimensions of the lightpipe with a micrometer and calculating die angles according to die previously described ray theory. The wavelength response was calibrated using various bandpass filters witii center wavelengths of 480 through 660 nm. A CCD (charge coupled display) camera with a macro lens was used to detect d e image on die white screen. A dark frame widi the lamp off was acquired widi the camera to account for thermal noise in the camera and stray background light. A nonresonant reference image to represent the lamp spectrum was acquired using TM light and a sample of air, which has an index too low to support SPR in this configuration. The dark frame was subtracted from both the reference and the sample images, and then the sample image was divided by the reference. For samples of acetone and water, the wavelength modulated spectra of d e six brightest angular bands coπesponding to die odd bands 13 dirough 23 were acquired using an image processing tool to analyze regions of me normalized image.

The reflection spectra obtained from the odd bands exiting the lightpipe were found to be extremely noisy due to die data acquisition method employed. The noise observed can be significantly decreased by imaging directly into a CCD camera. Six spectra were analyzed to find λ^ for each band. Altiiough the first order nature of the sensor makes it possible to utilize statistical calibration techniques (Martens, H. and Naes, T. (1989), Multivariate Calibration. John Wiley and Sons), a classical analysis of the location of the resonant wavelength for each angular band of the output was performed.

The resonance minima were determined by boxcar averaging of the collected experimental data to reduce noise, and then fitting a parabola to the data curve and mathematically determining the location of the minimum of the parabola. The sensor response was modeled witii the multilayer Fresnel reflection equations (Ishimaru, A. (1991) Electromagnetic Wave Propagation. Radiation, and Scattering. (Prentice Hall, New

Jersey), p. 43), using dispersive RI data for BK7 Schott Optical Glass Catalog. (1992), (Schott Glass Technologies, Inc., Duryea, PA) to represent the float glass light pipe, 302 A of silver (Gray, D.E. (Ed.) (1972), American Tnstimte nf Physics Handbook. 3rd ed., McGraw-Hill, New York, section 6, p. 149), 203 A of gold (Gray, D.E. (Ed.), American Institute of Physics Handbook. 3rd ed., McGraw-Hill, New York, section 6, p. 138), and bulk dielectric samples of acetone (Gray, D.E. (Ed.), American Tnstimte of Physics Handbook, 3rd ed., McGraw-Hill, New York, section 6, p. 105) and water (Palik, E.D. (Ed.) (1985), Handbook of Optical Constants nf Solids. Academic Press, Orlando, p. 1071). The measured center angles of each band are plotted against λ^,. in Figure 10 and compared to values calculated using die model. Figure 10 shows diat Δλ_sp-/ΔRI measured using die sensor closely matches values predicted by theory. For the 9.9° band the sensitivity is approximately 31 nm per 0.025 refractive index units (RIU) or, when inverted, δ.lxlO^-4 RlU/nm. The discrepancy between die model and die experimental data is primarily due to systematic eπor resulting from differences in die dispersive RI values used in die model calculation from their actual values (published RI values for BK7 were lower than those of the float glass of used for the experimental lightpipe substrate) and to die experimental uncertainty of the focus position on the center of the input face. There was also some uncertainty in determining die experimental values of λ_spr due to broad resonance curves and noise on die acquired spectra. Example 3:

A one-angle SPR symmetrical lightpipe sensor system, as shown in Figure 11, was fabricated using a standard 25 mm wide microscope slide as the lightpipe. The slide was cut to a length of L = 72.2 mm. Two metalized bevels were polished witii angles of α = 36° which coπesponds to an incidence angle θ = 72°. To support SPR, a 49.9 nm

(499A) thick gold layer was deposited on die top of a 2 nm (2θA) Cr adhesion underlayer to form the sensing area. The sensing area was 5 mm long (1 = 5 mm) and 10mm wide and positioned symmetrically along the length and width of die top surface of the lightpipe. A flow cell with a gasket seal was positioned over die metal sensing area on die top surface of the lightpipe was built to confine samples to the sensing area. Input light (white light) from a 100 watt halogen bulb (Oriel, Model 77501) was transmitted to the sensor by a 400 μm diameter optical fiber and collimated by a 10mm focal lengtii achromatic lens. The collimated light was TM polarized before it was coupled into the lightpipe. An identical collimator was used to collect the reflected light into an optical fiber and conducted it to die spectrograph detector.

Glycerol solutions of known concentrations were used as chemical samples in experiments to demonstrate die performance of the SPR one-angle lightpipe sensor. The SPR reflection spectra of these glycerol solutions were obtained using the planar lightpipe described above by dividing each output spectrum with a TE reference spectrum. The TE polarized component of the light does not excite SPR and serves as a simple way to record the spectrum of the lamp, fiber and detector. The normalized SPR reflection spectra for 4 different glycerol solutions are shown in Figure 12(a)-12(d).

The SPR resonance wavelength was determined from the SPR reflection spectrum by boxcar smoothing and parabolic curve fitting around d e reflection minimum of each spectrum. The calibration curve for SPR resonance wavelength and glycerol concentration by weight in solution was determined by least squares curve fitting of concentration of glycerol solutions versus SPR resonance wavelength. The calibration was found to be substantially linear over the concentration range examined (0.59%- 16.02% by weight), see Figure 13. Published values of the refractive indexes of aqueous glycerol solutions of different concentrations are shown on die alternate axis in Figure 13 to demonstrate the coπesponding RI calibration (Weast, R.C. (Ed.), (1985-1986) "CRC Handbook of Chemistry and Phvsics. 66th ed.,CRC Press, Inc., Boca Raton, Florida.

The stability of the one-angle SPR lightpipe sensor was assessed by measuring the fluctuation of the SPR resonance wavelength for each glycerol solution of Figure 13 over a 10 minute period. The short term test results, not shown, indicated excellent short term stability.

Longer stability experiments showed tiiat the standard deviation of the SPR resonance wavelength (σ^) over a two hour period is on the order of σ_spr = 0.1 nm. The standard deviation of the SPR resonance wavelength fluctuation over time can be used to predict die refractive index sensitivity of die lightpipe SPR sensor system. The local derivative of the refractive index calibration shown in Figure 13 was multiplied by σ_spr to find d e smallest resolvable change in RI (σ_n). The estimated value of σ_n for an RI around 1.345 is on the order of σ_n = 4x10 ^s. The same technique applied to die concentration calibration yields a sensitivity value of σ_conc(wt) of 3.4x10"* (weight %).

Example 4: A Multiple-Channel SPR Lightpipe Sensor

The beveled-ended zero order configuration of Figure 11 was designed to have a dual channel lightpipe sensor as shown in Figure 15. A standard 25 mm wide microscope slide was cut to a lengtii of 72.2 mm and divided lengdiwise into two channels adjacent to each other as shown in Figure 14A. A sensing area was fabricated on die top surface of die lightpipe by electron beam evaporation of a 2 nm thick Cr adhesion layer followed by a 50 nm thick gold SPR-supporting layer. The sensing area was 5 mm long and 20 mm wide. A flow cell widi a gasket (not shown) caπied samples to die sensing area. In this particular case, both sensing channels have identical sensing areas. Output from the two different sensing channels, at different positions along the width of the lightpipe, exited die lightpipe at coπesponding positions along d e width of d e lightpipe and was separately collected and analyzed in a detector system. White input light from a 100 W halogen bulb (Oriel , Model 77501) was TM polarized and transmitted to die lightpipe by an optical fiber. Input light was then collimated using a 15 mm focal length achromatic lens and coupled in to die lightpipe dirough the bottom surface by reflection off of the beveled input end (α = 36°). The input light was shared by both channels, i.e. distributed across the widdi of the lightpipe. In this case, the diameter of the lens was selected so that a sufficiently large widdi of die input face was illuminated to allow input light to interact with both sensing channels.

Both sensing channels were contacted with the same samples, a series of aqueous glycerol solutions ranging in concentration from 0.0% - 30% (by weight), and both sensing channels were maintained at constant temperamre, 25°C. The literature values for the RI's of the samples ranged from 1.334 to 1.369.

If the input beam is not perfectly collimated, light at a range of incident angles interacts with and excites SPR in the sensing area. An input beam can be uncoUimated to a given consistent degree by offsetting an input fiber from the focal point of the lens. As in the lightpipe sensor of Figure 3 A, the output is a set of angular bands, and different angular bands contain SPR spectra excited at different incident angles.

To demonstrate that two different SPR spectra can be obtained from adjacent channels on the lightpipe surface, the SPR spectra from the two channels for two different incidence angles (i.e., of two different angular bands) were measured. The SPR reflection spectra of the aqueous glycerol solutions were obtained by channel selective detection of the output spectra of different angular ouφut bands and dividing each ouφut spectrum by the TE polarized component of d e light. (This coπects for the spectrum of the light, fiber and detector.) The SPR wavelength (λ_SPR) was determined from these reflection spectra by boxcar smoothing and parabolic curve fitting around die reflection minimum of each spectrum. The relationship between λ_SPR and glycerol concentration of both channels was determined by using a least square curve fitting as shown in Figure 16A. The derivatives of the concentration curve (δλs_PR/δconc) are plotted as a function of glycerol concentration in Figure 16B to compare the SPR signal shifts due to die refractive index variations in d e spectra from two channels (different incidence angles). The concentration calibration curves of the two channels are different. This confirms that die ouφut of the two channels can be separately measured and analyzed. Analogously, SPR signal shifts derived from the ouφut of the two different channels from the same angular band are substantially the same (not shown). Consequently, in an analogous dual- channel lighφipe sensor, if one of the sensing areas is activated and the other is not, the non-activated sensing area can be used as a source of a reference signal to compensate for system variables including temperamre variation.

Those of ordinary skill in the art will appreciate that methods, materials and techniques otiier than those specifically discussed herein can be readily employed or adapted to implement the sensor configurations and practice die methods of this invention. For example, a variety of means for measuring reflection coefficients and/or light intensity, particular as a function of wavelength, are well-known and available to those in the art. In addition, there are a variety of techniques and devices known for collimating, collecting, focusing and conducting light that can be applied or readily adapted to light input to or light ouφut from the sensors of this invention. Those of ordinary skill in the art can readily select from among such alternatives, variants and functional equivalents those that are appropriate for use in the SPR configuration of this invention.

All of the references cited in this specification are incorporated by reference in their entireties herein.

Claims

WE CLAIM:

1. A surface plasmon resonance sensor which comprises:

a planar lighφipe having an input end, an ouφut end and a sensing area on an external planar surface, said sensing area comprising a conducting layer tiiat supports surface plasmon resonance;

a light source optically connected to the input end of said lighφipe to introduce light into said lighφipe at a range of angles such tiiat the light is conducted through said lighφipe by total internal reflection to reflect off said sensing area, exciting a surface plasmon wave therein, and to exit said lighφipe at said ouφut end in angular bands;

a detector for receiving angular bands of light exiting said ouφut end and reflected from said sensing area which thereby detects surface plasmon resonance.

2. The sensor of claim 1 wherein said light source is a multiple wavelength light source.

3. The sensor of claim 1 wherein light is introduced into said lighφipe by focusing said light at the input end of the lighφipe.

4. The sensor of claim 1 wherein said light is introduced asymmetrically at a range of angles filling angles on only one side of d e optical axis of die lighφipe.

5. The sensor of claim 1 wherein said light is introduced into said lighφipe at a range of angles filling angles on eidier side of die optical axis of the lighφipe.

6. The sensor of claim 1 wherein said lighφipe comprises two or more sensing areas on a planar surface.

7. The sensor of claim 6 wherein one of said sensing areas is a reference sensing area.

8. The sensor of claim 6 wherein said lighφipe comprises a plurality of sensing areas across the width of the lighφipe.

9. The sensor of claim 6 wherein said lighφipe comprises a plurality of sensing areas along the length of the lighφipe.

10. The sensor of claim 1 wherein said sensing area comprises a SPR-supporting metal layer.

11. The sensor of claim 10 wherein said sensing area comprises an adherence layer.

12. The sensor of claim 10 wherein said sensing area comprises a reactive layer.

13. The sensor of claim 12 comprising a plurality of sensing areas along the width of die lighφipe and wherein at least one of said sensing areas is a reference sensing area.

14. The sensor of claim 13 wherein said sensing areas that are not reference sensing areas each comprise a reactive layer.

15. The sensor of claim 14 wherein in each of said sensing areas comprising a reactive layer, the reactive layer is specific for a different analyte.

16. The sensor of claim 12 wherein the sensor is a biosensor.

17. The sensor of claim 1 wherein the detector is a spectrograph.

18. The sensor of claim 1 wherein the detector is a CCD camera.

19. The sensor of claim 1 wherein the detector comprises a fiber optic pickup.

20. The sensor of claim 1 which comprises a telecentric lens optically coupled to the ouφut end of said lighφipe.

21. The sensor of claim 20 wherein said telecentric lens is a cylindrical telecentric lens.

22. The sensor of claim 21 wherein said telecentric lens is a spherical telecentric lens.

23. The sensor of claim 21 wherein angular bands of light exiting said lighφipe are redirected and focused by said telecentric lens to be substantially parallel to die optic axis of die lens.

24. The sensor of claim 1 wherein d e input end of the lighφipe is beveled.

25. The sensor of claim 24 wherein input light is focused at the external face of the beveled input end.

26. The sensor of claim 25 wherein both the input and ouφut end of the lighφipe are beveled.

27. The sensor of claim 26 wherein the reflective surface of the bevel comprises a diffraction grating.

28. The sensor of claim 24 wherein said input and ouφut ends are beveled at 45° with respect to the normal to the bottom surface of the lighφipe.

29. The sensor of claim 28 wherein input light is focused at the internal face of said beveled input end through the bottom planar surface of the lighφipe.

30. The sensor of claim 1 further comprising a sample cell adjacent to said lighφipe sensor which allows a gas or liquid sample to interface with a sensing area of the lighφipe.

31. The sensor of claim 30 wherein said sample cell is a flow cell.

32. The sensor of claim 1 further comprising a TM polarizer optically coupled anywhere witiiin said sensor to exclude TE polarized light from said detector.

33. The sensor of claim 1 further comprising a first cylindrical lens optically coupled between said light source and the input face of said lighφipe which focuses light passing diere through at the input face of the lighφipe at a range of angles.

34. The sensor of claim 33 further comprising a second cylindrical lens optically coupled between said light source and die said first cylindrical lens, wherein said second cylindrical lens collimates light across width of the lighφipe.

35. The sensor of claim 1 wherein said lighφipe is fabricated from glass or crystal.

36. The sensor of claim 1 wherein said lighφipe is substantially coated witii a cladding layer having an index of refraction different from that of the lighφipe substrate except that the cladding does not cover said sensing area.

37. The sensor of claim 1 wherein said lighφipe is fabricated from plastic or a polymer material.

38. The sensor of claim 37 wherein said lighφipe is substantially coated witii a cladding layer having an index of refraction different from that of the lighφipe substrate except that the cladding does not cover the sensing area.

39. A method for surface plasmon resonance measurement of a sample, which method comprises the steps of contacting the sample with a sensing area of a surface plasmon resonance lighφipe sensor of claim 1 and detecting the reflection spectrum of an angular band of light exiting said lighφipe.

40. A mediod for surface plasmon resonance measurement of a plurality of analytes in a sample, which method comprises the steps of contacting the sensing areas of an

SPR lighφipe sensor of claim 8 with die sample and detecting the reflection spectrum of angular bands of light exiting said lighφipe.

41. A mediod for surface plasmon resonance measurement of a plurality of analytes in a sample, which method comprises the steps of contacting the sensing areas of an

SPR lighφipe sensor of claim 9 with the sample and detecting the reflection spectrum of angular bands of light exiting said lighφipe.

42. A planar lighφipe having a plurality of SPR-supporting sensing areas along d e length and widdi of an external planar surface diereof.

43. The planar lighφipe of claim 42 wherein said SPR-supporting sensing areas comprise an SPR-supporting metal layer.

44. The planar lighφipe of claim 43 wherein one or more of said SPR-supporting sensing areas further comprise a reactive layer.

45. The planar lighφipe of claim 42 wherein at least one of said SPR-supporting sensing areas is a reference sensing area.

46. A surface plasmon resonance sensor which comprises:

a planar lighφipe having a beveled input end, a beveled ouφut end and a sensing area on an external planar surface said sensing area comprising a conducting layer that supports surface plasmon resonance; a light source optically connected to the input end of said lighφipe to couple light into said lighφipe at a single angle by reflection off the bevel of said input end such that light is conducted dirough said lighφipe by total internal reflection to reflect off said sensing area exciting a surface plasmon wave therein and to exit said lighφipe by reflection off said ouφut bevel; and

a detector for receiving ouφut light exiting said ouφut end of die lighφipe and measuring spectral ouφut of said lighφipe as a function of wavelength which thereby detects surface plasmon resonance.

47. The sensor of claim 46 wherein said lighφipe comprises two or more sensing areas on a planar surface.

48. The sensor of claim 47 wherein one of said sensing areas is a reference sensing area.

49. The sensor of claim 47 wherein said lighφipe comprises a plurality of sensing areas across die widdi of die lighφipe.

50. The sensor of claim 46 wherein said sensing area comprises a SPR-supporting metal layer.

51. The sensor of claim 50 wherein said sensing area comprises an adherence layer.

52. The sensor of claim 50 wherein said sensing area comprises a reactive layer.

53. The sensor of claim 49 comprising a plurality of sensing areas along the widdi of die lighφipe and wherein at least one of said sensing areas is a reference sensing area.

54. The sensor of claim 53 wherein said sensing areas that are not reference sensing areas each comprise a reactive layer.

55. The sensor of claim 54 wherein, in each of said sensing areas diat comprises a reactive layer, die reactive layer is specific for a different analyte.

56. The sensor of claim 46 wherein the sensor is a biosensor.

57. The sensor of claim 46 wherein the detector is a spectrograph.

58. The sensor of claim 46 wherein the detector comprises a fiber optic pickup.

59. The sensor of claim 46 wherein the external surface of both beveled ends are coating with a reflective surface.

60. The sensor of claim 46 wherein both ends are beveled at an angle of 36° widi respect to the bottom surface of the lighφipe.

61. The sensor of claim 46 further comprising a sample cell adjacent to said lighφipe sensor which allows a gas or liquid sample to interface with a sensing area of the lighφipe.

62. The sensor of claim 61 wherein said sample cell is a flow cell.

63. The sensor of claim 46 further comprising a TM polarizer optically coupled in the system to remove TE polarized light before light enters the detector.

64. The sensor of claim 46 further comprising a collimator optically coupled to said light source and a TM polarizer optically coupled between said collimator and said input end of said lighφipe.

65. The sensor of claim 46 wherein light is coupled into said lighφipe through said bottom planar surface of said lighφipe substantially normal to said surface.

66. The sensor of claim 46 wherein said sensing area is located in the center of the top surface of said lighφipe.

67. The sensor of claim 46 wherein said lighφipe is fabricated from glass.

68. The sensor of claim 46 wherein said lighφipe is substantially coated witii a cladding layer having an index of refraction different from that of the lighφipe substrate except that the cladding does not cover said sensing area.

69. The sensor of claim 46 wherein said lighφipe is fabricated from plastic.

70. The sensor of claim 69 wherein said lighφipe is substantially coated witii a cladding layer having an index of refraction different from that of the lighφipe substrate except that the cladding does not cover the sensing area.