WO1995001554A1 - Solid-phase binding assay system and method for interferometrically measuring analytes bound to an active receptor - Google Patents

Abstract

Apparatus (10) and method for detecting and measuring analyte in biological and other sample solutions. A substrate (20) is provided with at least one active region which specifically binds analyte and an inactive region that does not, and a change in optical path length through the active region due to binding of analyte is measured. The substrate (20) is inserted into one or both beams of an interferometer (10) and moved relative to the interferometer (10). As the substrate (20) moves, the active and inactive regions pass successively through the beam(s) causing a periodic phase shift in the light. When the beams, one or both of which have undergone phase modulation, are recombined with each other, the phase modulation is converted into an amplitude modulation. Each recombined beam is directed to a photodetector (42, 45), which converts the periodically varying optical power into a periodically varying electrical signal. This signal has an amplitude proportional to the amount of bound analyte on the surface of the substrate (20) and a signal frequency equal to that at which the active regions move past the beam(s) in the interferometer (10). Servo control (60, 62, 65) is used to maintain the interferometer at a desired operating point that maximizes the sensitivity and linearity of the system so that the small phase differences due to analyte binding can be measured more accurately.

Description

SOLID-PHASE BINDING ASSAY SYSTEM AND METHOD FOR INTERFEROMETRICALLY MEASURING ANALYTES BOUND TO AN ACTIVE RECEPTOR

CROSS REFERENCE TO RELATED APPLICATION

This application is a continuation-in-part of application No. 08/086,345, filed June 29, 1993, the entire disclosure of which is incorporated by reference.

BACKGROUND OF THE INVENTION

This invention relates to a solid-phase binding assay system for measuring analytes.

Many solid-phase immunoassays involve surface illumination and consequent light emissions from molecules attached to the surface. Generally, these emissions travel in all directions. Either these divergent emissions must be collected with expensive and awkward light collection optics to achieve high sensitivity or the inherent inefficiencies and consequent low signal-to-noise ratio must be accepted. Many immunoassay systems have been developed using different physically measurable properties of reagents to provide a measurement of an analyte concentration. Radio immunoassay (RIA) , immunofluorescence, chemiluminescence, enzyme immunoassays (EIA) , free radical immunoassays (FRAT) , light scattering nephelometry, transistor bridge probes, indium reflective surfaces, and ultrasonic probes have been applied. These systems use the highly selective reaction between a primary binding reagent material such as an antibody or antigen and an analyte selectively binding therewith. An attempt by others to develop an optical probe comprising a metal covered diffraction grating coated with a monoclonal antibody specific for a virus, bacterium or other desired antigen has been described by Moffatt, A. Genetic Engineering News, p. 18, October 1986. The shift in wavelength of reflected light is apparently determined and correlated to a concentration in analyte.

A reflectance method for quantification of immunological reactions on polished crystalline silicon wafer surfaces has been described by Arwin, H. et al, Analytical Biochemistry, 154:106-112 (1985). Indium surface reflection methods are described by Giaever in U.S. Pat. Nos. 3,853,467, 3,926,564, 3,960,488, 3,960,489, 3,960,490, 3,975,238, 3,979,184, 3,979,509, 4,011,308, 4,018,886, 4,054,646, 4,115,535, 4,172,827 and 4,181,501. Liquid layer thicknesses can be monitored by a reflectance method described in U.S. Patent No. 3,960,451.

Biosensors such as field effect transistors probes and their use in assays are described by Pace, S., Medical Instrumentation, (19(4) :168-172 (1985). Polysilicon surfaces are included among the possible biosensor surfaces to which primary binding reagents can be attached.

Other patents of interest include U.S. Pat. Nos. 4,537,861; 4,558,012; 4,647,544; 4,820,649; 4,876,208; 4,886,761; 5,089,387; 5,120,131; 5,196,350; and RE 33,581.

The most sensitive instrument for the measurement of optical phase shift due to physical displacement or refractive index change or both is the optical interferometer. There are many different interferometer designs. Measurements made with interferometers include: 1) the attempt to detect gravitational radiation (A. Abramovici, et al. "LIGO: The Laser Interferometer Gravitational-Wave Observatory" Science 256, April 17, 1992, p. 325), 2) positioning of the cutting head and work piece in precision machining (C. Evans, "Precision Engineering: an Evolutionary View", Cranfield Press 1989), 3) tectonic plate movement in geology for earthquake prediction (P.L. Bender, "Laser Measurement of Long Distances", Proc. IEEE, Vol. 55, No. 6, June 1967, PP. 1039-1045) and 4) icroarcsecond astrometry (R.D. Reasenberg, et al. "Microarcsecond Optical Astrometry: An Instrument and its

Astrophysical Applications", The Astronomical Journal, 96 (5) 1988) . SUMMARY OF THE INVENTION The present invention provides assays and apparatus having improved sensitivity by which it is possible to detect and measure analytes present at very low concentrations in biological and other sample solutions. Assays and apparatus of the present invention are characterized by high linearity and do not require labeling, although labeling may be advantageously employed under certain circumstances. Further, the assays of the present invention have a large dynamic range, i.e., are capable of measuring over many orders of magnitude of analytic concentration without requiring reagent concentration changes. In other words for a given test only one analyte- binding substrate (as described below) should be necessary. Additionally the assay produces a background free output such that when no analyte exists in the incubating solution, very little or no signal results. The technique involves few process steps, can be used with short incubation times, and has low coefficient of variation in relation to a standard measurement. According to a broad aspect of the present invention, a substrate is provided with at least one active region which specifically binds analyte, e.g., antigens, antibodies, or other target substances that can be specifically bound to an active region, and a change in optical path length through the region due to binding of analyte is measured. This is preferably done by comparing the phase of light that has passed through the active region with the phase of light that has not. The phase comparison is preferably performed with an interferometer. One set of embodiments includes a moving substrate having the active region(s) formed on -., surface thereof, usually being disposed in an alternating pattern with inactive region(s) , i.e., area(s) on the substrate surface which do not bind analyte. In embodiments where the substrate is a disc that is rotated, preferred patterns for the active and inactive regions include spots and wedges.

The interferometer splits an input beam (typically from a laser) into two beams, which are directed along two paths, and subsequently recombined to provide one or two recombined beams. The interferometer may be any of a number of types, including Mach-Zehnder and Michelson interferometers. The substrate is inserted into one or both beams of the interferometer and moved relative to the interferometer. As the substrate moves, the active and inactive regions pass successively through the beam(s) causing a periodic phase shift in the light. When the beams, one or both of which have undergone phase modulation, are recombined with each other, the phase modulation is converted into an amplitude modulation.

Each recombined beam is directed to a photodetector, which converts the periodically varying optical power into a periodically varying electrical signal. This signal has an amplitude proportional to the amount of bound analyte on the surface of the substrate and a signal frequency equal to that at which the active regions move past the beam(s) in the interferometer.

While placing a moving object inside an interferometer is contrary to normal practice, the advantages can be made to outweigh the disadvantages. In particular, the signal component that represents the small phase changes due to analyte binding is an AC signal. Since the spectrum of disturbances to the interferometer decreases with frequency, most of the noise can be avoided by suitable placement of the AC signal (say at several kilohertz) , and more precise measurements can be made. For example, if the substrate is moving rapidly and there are many active regions (e.g., spots or wedges around a disc) , the signal frequency will be much higher than the frequency of the noise due to substrate wobble, variation in the substrate thickness, and vibrations imparted to the interferometer by the motor moving the substrate.

In certain embodiments, servo control is used to maintain the interferometer at an operating point where the phases of the two beams in the interferometer differ by 90° (plus integral multiples of 180°). This maximizes the sensitivity and linearity of the system so that the small phase differences due to analyte binding can be measured more accurately. In a particularly preferred aspect of the present invention, the inactive region(s) on the substrate surface are treated so that they will non-specifically bind analyte and non-specifically bind substances other than analyte in amounts similar to that which the active region non-specifically binds such substances. As the active region(s) will also be expected to non-specifically bind substances other than analyte, the amount or thickness of substances non-specifically bound within the inactive region(s) will offset or correct inaccuracies resulting from binding of substances other than analyte within the active region(s) , when the substrate is read in the preferred two-beam interferometer.

In other embodiments, the substrate would not be moved, but the beam would be expanded so that a large number of active regions on the substrate could be interrogated at one time.

A further understanding of the nature and advantages of the present invention may be realized by reference to the remaining portions of the specification and the drawings.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic diagram of an embodiment of the system of the present invention, using a Mach-Zehnder interferometer; Fig. 2 is a plan view of one embodiment of a disposable sample disc employed in the system of the present invention;

Fig. 3 is an idealized plot of phase difference due to analyte binding as a function of position on the disk; Fig. 4 is a graph showing detected optical power as a function of phase difference and illustrating a preferred operating point;

Fig. 5 is a block diagram of analog circuitry for performing phase-sensitive detection; Fig. 6 is a block diagram of circuitry for digital data analysis; Fig. 7 is an optics diagram illustrating the optical geometry of a beam passing through the sample disc of the present invention;

Fig. 8 is a schematic diagram of a modified version of the embodiment of Fig. 1;

Fig. 9 is a schematic diagram of a combined beamsplitter-mirror element;

Fig. 10 is a schematic block diagram of a Mach- Zehnder interferometer using the element of Fig. 9; Fig. 11 is a schematic diagram of a variant of the element of Fig. 9;

Fig. 12 is a schematic diagram of an alternative interferometer embodiment;

Fig. 13 is a schematic diagram of a further alternative interferometer embodiment;

Fig. 14A and 14B are schematic diagrams of Michelson interferometer embodiments using reflection;

Fig. 15 is a plan view of an embodiment of a disposable sample disc having spots on a spiral path; Fig. 16 is a plan view of an embodiment of a disposable sample disc having wedge-shaped active and inactive regions;

Figs. "17A-17H" are plan views showing different arrangements of active and inactive binding regions on the disc and beam positions relative to the disc;

Figs. 18A-18H are idealized plots of phase difference due to analyte binding for the disc and beam geometries of Figs. 17A-17H;

Fig. 19 is a plot of interferometrically-measured thickness as a function of analyte concentration in a competitive biotin binding assay;

Fig. 20 is a plot of interferometrically-measured thickness as a function of analyte concentration in a CKMB assay; Fig. 21 is a plot of the interferometer voltage signal, taken from the differential amplifier output for an actual disc sample; and Fig. 22 is a plot of the interferometer voltage signal, taken from the servo amplifier output for an actual disc sample.

DESCRIPTION OF SPECIFIC EMBODIMENTS

I. Biological Overview

The present invention is useful in assaying for a wide variety of analytes in virtually any type of sample which is liquid, which can be liquified, or which can be suspended in a liquid. The method and apparatus will find their greatest use with biological specimens, such as blood, serum, plasma, urine, cerebral fluid, spinal fluid, ocular lens liquid (tears) , saliva, sputum, semen, cervical mucus, scrapings, swab samples, and the like. Use will also be found with industrial, environmental and food samples, such as water, process streams, milk, meat, poultry, fish, conditioned media, and the like. Under certain circumstances, it may be desirable to pretreat the sample, such as by liquefaction, separation, dilution, concentration, filtration, chemical treatment, or a combination thereof, in order to improve the compatibility of the sample, with the remaining steps of the assay. The selection and pretreatment of biological, industrial, and environmental samples prior to immunological testing is well known in the art and need not be described further. The analyte to be detected may be virtually any compound, composition, aggregation, or other substance which may be immunologically or otherwise specifically bound within an active region on a substrate, as described in detail below. That is, the analyte, or portion thereof, will usually be antigenic or haptenic having at least one determinant site, or will be a member of a naturally-occurring binding pair, e.g., enzyme and substrate, carbohydrate and lectin, hormone and receptor, complementary nucleic acids, and the like. Analytes of particular interest include antigens, antibodies, proteins, carbohydrates, haptens, drugs, hormones, hormone metabolites, macromolecules, toxins, bacteria, viruses, enzymes, tumor markers, nucleic acids, and the like, although other types of substances may also be detected. A non-exhaustive list of exemplary analytes is set forth in U.S. Patent No. 4,366,241, at column 19, line 7 through column 26, line 42, the disclosure of which is incorporated herein by reference.

Presence of the analyte in the sample will be measured by detecting specific binding between the analyte and an anti-analyte covalently or non-covalently bound to a surface on a solid phase substrate, usually a disc as described in detail below. The anti-analyte will be a specific binding substance capable of binding directly or indirectly to the analyte (or labelled analyte analog) with a high affinity, typically being at least about 10⁸ M^"1, usually being at least about 10⁹ M^"1, and sometimes being 10¹⁰ M^"1 or greater. The anti-analyte should be free from cross-reactivity with other substances that may be present in the sample or the assay reagents. Most commonly, the anti-analyte will be a monoclonal or polyclonal antibody raised against the analyte, but in some cases it may be possible to employ natural receptors for biological analytes, as described above. In cases where the analyte is itself an antibody, it will of course be possible to employ antigens or haptens recognized by the antibody as the anti-analyte.

While the anti-analyte will most often bind directly to the analyte, the present invention also comprises indirect binding of anti-analyte to analyte, i.e., the use of one or more intermediate binding substances to sequester or effect a linkage to the analyte. For example, when binding to a solid phase substrate, it will be possible to provide a primary binding substance, e.g., avidin or a primary antibody, on the solid phase which is able to bind directly a soluble substance which is specific for the analyte, e.g., a biotinylated antibody or secondary antibody which recognizes the analyte. A wide variety of such indirect binding protocols are available and well described in the scientific and patent literature. The term "anti-analyte" as used in the specification and claims are thus intended to include all substances which are able to bind the analyte , either directly (i.e., without an intermediate binding substance) or indirectly (i.e., with one or more intermediate binding substances forming a linkage) . The apparatus and methods of the present invention will often advantageously employ amplified binding systems where capture of a single analyte molecule on the solid phase substrate surface will result in binding of at least one additional molecule, particle, or the like, to increase the total amount and/or thickness, of material which is bound to the analyte-specific portion of the analyte-specific portion of the analyte receptor region. In this way, the sensitivity of analyte detection can be increased without any need to modify the system hardware. Suitable immunological amplification techniques are well described in the patent and scientific literature. See, for example, U.S. Patent Nos. 4,463,090; 4,687,736; and 4,595,655, the full disclosures of which are incorporated herein by reference. Other preferred techniques of the present invention will provide for binding of a dense marker particle, such as a gold particle, to the analyte- specific portion of the analyte receptor region. Such dense particles will significantly enhance detectability using the interferometer systems described hereinafter. The term "binding assay", is used herein to designate an assay using any binding reaction between a binding reagent and the other member of the binding pair which is selectively bindable therewith.

II. System Overview a. Instrument Configuration

Fig. 1 is an optical and electronic schematic showing the basic elements of an embodiment of an interferometric immunosensor system 10 according to the present invention. In brief, the system includes an interferometer subsystem 12 and an electronic subsystem 15. A disposable substrate 20 has active regions to which target analyte (if present in a sample) has been bound. The active regions are normally separated by inactive regions, namely regions to which the target analyte does not bind. Bound analyte increases the optical thickness of the substrate in the active regions. As will be described in greater detail below, the inactive regions can be treated so as to non-specifically bind substances in amounts equal to the active regions and thus compensate for the problem of non¬ specific binding.

The substrate is incubated and then interposed in at least one of the light paths of the interferometer and, in most embodiments, moved relative to the light paths, which undergo phase modulation in accordance with the amount of bound analyte. The electronic subsystem receives output signals from the interferometer subsystem and provides a signal representing the amount of bound analyte. The electronic subsystem also preferably provides active control of the interferometer to maintain certain operating conditions.

Interferometers operate by comparing the phase difference between two light beams that have originated from a single source and have been separated to travel over different paths. The difference in phase is then read out by recombining the beams, thus converting this phase difference to an amplitude difference.

Fig. 1 shows a Mach-Zehnder interferometer, which includes a laser 30, a first section including a first 50/50 beamsplitter 32 and a first mirror 35, a second section including a second mirror 37 and a second 50/50 beamsplitter 40, and photoreceivers 42 and 45. Mirror 37 is mounted to a fixed mirror mount 50 so as to be fixed relative to the other elements of the interferometer, while mirror 35 is mounted for controlled relative movement. To this end, mirror 35 is mounted to a piezoelectric transducer 52, which is itself mounted to a fixed mirror mount 55. Laser 30 can be any convenient type (e.g., a helium-neon laser or a diode laser operating in the visible or near infrared) . The beamsplitters are shown schematically as dashed lines, and can be of any desired construction (e.g., coated optical flats, prism cubes, or fiber optic couplers) .

The electronic subsystem includes a differential amplifier 60, a servo amplifier 62, a high-voltage amplifier 65, and signal processing electronics 67, which may be analog or digital.

In the specific embodiment, substrate 20 is a disc, which is interposed in both beams and rotated by a motor 70. While it is also possible to practice the invention by translating the substrate, rotation is preferred since embodiments that rotate the substrate tend to be simpler, cheaper, and quieter. The substrate will often be referred to as the disposable disc, or simply the disc. A position encoder 72 is coupled to motor 70 and provides a reference signal to the signal processing electronics. If there are N active regions disposed at equal intervals around the perimeter of the disc and the disc is rotated at a rotational frequency of Ω, the beam(s) will undergo phase modulation at a signal frequency, f_Si_g, equal to the product of N and Ω. That is, f_sig=NΩ. The frequency of the reference signal from encoder 72 is preferably at the same frequency f_sig or at an integer multiple of f_{S g}. In the Mach-Zehnder geometry the light output from laser 30 encounters beamsplitter 32 where it is split into two equal power beams, referred to as the reference beam (which is reflected by the beamsplitter) and the signal beam (which passes through the beamsplitter) . The designation of the beams as the signal and reference beams has significance for those embodiments where only one beam is modulated by the active regions on the substrate. In many embodiments, both beams are modulated, so the designation of one beam as the signal beam and the other as the reference beam is somewhat arbitrary. The reference beam is reflected from mirror 35 and then passes through disc 20 to beamsplitter 40. The signal beam passes through the spinning disposable disc and is then reflected from mirror 37 to beamsplitter 40, where it is recombined with the reference beam. The recombined reference and signal beams are passed to photoreceivers 42 and 45, the outputs of which are applied to differential amplifier 60.

The output of differential amplifier 60 is applied to servo amplifier 62, the output of which is applied to high- voltage amplifier 65, which drives PZT 52. As will be described below, it is preferred to use PZT 52 to move mirror 35 to maintain a position where the path length difference between the two beams corresponds to a phase shift where exactly half of the light falls on photoreceiver 42 and half falls on photoreceiver 45.

The figure shows the input to siσnal processing electronics 67 drawn in dashed lines connec ed to the output of the differential amplifier (point A) and the output of the servo amplifier (point B) . As will be described below, the point from which the signal that represents bound analyte is measured depends on the servo bandwidth. This signal, which is input to signal processing electronics 67, will sometimes be referred to as the analyte signal.

b. Disc Configuration and Operation Overview Fig. 2 shows in plan view two rotational positions of a disc-shaped substrate 20 having active spots 20a of anti- analyte alternating with inactive spots 20b of a non-specific binding substance around the outer portion of the disc surface. While the figure shows only eight of each type of spot, a significantly larger number of spots is preferred, usually at least 16, preferably at least 128, and often as many as 512, or more.

The figure also shows a pair of X*s denoting one possible arrangement of the signal and reference beams. In this arrangement one beam is on an active spot while the other beam is on an adjacent inactive spot. When the disc rotates by an amount equal to the interval between adjacent spots (22.5° in the illustrated example) , the beam that was passing through an active spot passes through an inactive spot (and vice versa) .

Fig. 3 is a plot of the phase difference due to analyte binding as a function of angular position (or time) as the incubated disc is rotated. As will be discussed below, these phase differences can be very small compared to the other phase disturbances in the system (especially if labeling is not used) , but can be detected accurately. The plot is an idealization since it is assumed that (1) the spots have the same area and the same spacing, (2) all the active and inactive spots have the same optical thickness before incubation, (3) all the active spots have the same optical thickness after incubation, (4) all the inactive spots have the same optical thickness after incubation, (5) the beam area is much smaller than the spot area, and (6) the disc is flat, the faces are parallel and the index of refraction is uniform. If the phase differences were measured prior to incubation, there would be no phase differences due to the spots, and the plot would be a horizontal line through the origin. The reasons for this are as follows. When the spots were in the beam paths they would produce no phase difference since the active and inactive spots have the same optical thicknesses. As the spots were moving out of the beam paths, the transitions would be negligible.

After incubation, however, the active and inactive spots would differ in optical thickness, and would produce phase differences of alternating sign, as shown. The phase difference would be zero during the intervals that the spots were out of the beam paths.

c. Photoreceiver and Servo Overview Photoreceivers 42 and 45 provide voltage outputs representing optical power. Specifically, each photoreceiver includes a photodiode, a transimpedance amplifier, and a voltage amplifier. The photodiode converts optical power P into a photocurrent i, which is related to the optical power by the current responsivity R, as follows:

where e is the electron charge, η is the quantum efficiency, h is Planck's constant, and . is the laser frequency. The photodiode current is converted into a voltage V_λ by the transimpedance amplifier with a feedback resistor R_τ as follows:

V₁ = i The voltage V after the photoreceiver gain stage with gain G_rec is given by:

^{V = G}rec^Rτi ^{= G}rec^RT^RP As will be shown in detail below, the optical powers P₄₂ and P₄₅ on photoreceivers are given by:

P_Λ2 = ^[l-cos(Δ)]

P>_ = -^[l+cos(Δ)]

where Δ is the phase difference between the two paths.

The two voltages from the two photoreceivers are amplified by differential amplifier 60 whose voltage output is equal to the difference between the two photoreceiver voltages. The output from the differential amplifier is called the error signal V_error and is given by:

V_βrror = G_xβcR→jR [P₄₅ - P₄₂] = G^R-pRP-COS (Δ )

Servo amplifier 62 provides gain and phase control and controls the bandwidth of the error signal. High-voltage amplifier 65 amplifies the error signal up to the high voltage (in some cases a few hundred volts) required to move the PZT. The response of the interferometer to the small phase shifts produced by the bound analyte is a function of the difference between the two paths through the interferometer. The output from the high-voltage amplifier is called the control signal, and is applied to PZT 52 to maintain the phase difference between the two paths through the interferometer at a desired operating point so as to achieve maximum sensitivity and linearity.

Fig. 4 is a plot of the individual powers on the two photodetectors, and the difference between the powers, as a function of phase difference Δ (with the servo turned off) .

The value of Δ where the signal from the photodetectors changes most rapidly with changes in Δ is π/2 (90°), or more generally (n+_)π where n is an integer, and is called the operating point of the interferometer. At the operating point the absolute value of the derivative of the signal power with respect to the phase difference is a maximum, and the sensitivity is therefore also a maximum.

The error signal, which is proportional to the difference between the powers on the two photoreceivers, has several excellent properties for feedback control. In particular, the slope is steepest at the operating point, and the signal is highly linear near and around the operating point. Further, the signal is signed, i.e., when the phase difference to be controlled is zero, the error signal is zero, and the error signal changes sign from positive to negative as it goes through zero.

By way of an example of a condition to be avoided, consider a condition where the phase difference Δ between the two paths is zero. The derivative of the signal power with respect to Δ would also be zero, and the interferometer would be insensitive to the small phase changes that are sought to be measured.

The interferometer is subject to environmental disturbances that can change the path lengths. Additionally, the spinning disc can produce phase disturbances that will move the interferometer away from the operating point. This results from such effects as disc wobble and variations in thickness in the disc from point to point. These phase disturbances must be controlled by feedback if the interferometer is to remain at the operating point.

d. Signal and Noise Overview

The present invention operates to extract small signals from a very noisy background. Although some of the specific types of noise and techniques for dealing with them will be discussed in detail below, some order-of-magnitude figures are set forth here to provide a context for the remaining discussions. A typical size of antibodies that might be used as the binding substance on the active regions is on the order of 60 A (0.006 μm) when the antibodies are on end with their active portions pointing away from the disc. The antibodies do not cover the entire disc surface. The size of antigens that might be the analyte can vary widely from a few to several hundred angstroms. As a result, the bound layer may have a thickness only on the order of 100 A. By way of reference, the wavelength of a helium-neon laser is 6328 A (0.6328 μm) in air. Labels may be used to increase the thickness of the bound layers, and should be as large as is consistent with efficient incubation. While large particles provide large phase shifts, they diffuse slowly and therefore very few of them may reach the disc surface. Smaller labels produce smaller phase shifts, but many more of them may reach the surface. Typical labels have diameters in the range of 500- 10,000 A.

Some sources of noise produce phase shifts that can be much larger than the phase shifts produced by the bound analyte. For thickness variations that have a spatial frequency commensurate with the disc diameter, the phase shifts can be on the order of radians. Fortunately, the servo will cancel out most of the resultant phase differences since these variations will manifest themselves as noise at or near the rotational frequency, which is well within the bandwidth of the servo system. Some thickness variations have a higher spatial frequency, perhaps commensurate with that of the active regions on the disc. The RMS roughness at the measurement spatial frequency decreases as the spatial frequency increases, but can be as large as tens of angstroms, and cannot be removed by the servo.

The noise due to vibrations of the optical components can be significant at the resonant frequency of those components. For example, in a specific prototype, mirror mounts exhibited a resonance at 800 Hz, and could be stimulated by sound at that frequency to vibrate with an amplitude of several tens of angstroms. Fortunately, it is usually possible to place the signal frequency at a value that is far from any resonant frequencies of the components.

Noise in the detectors and electronics can also be significant. Again, much of the random noise is at frequencies below the signal frequency. III. Signal Processing a. Phase-Sensitive Detection

Fig. 5 is a block diagram of one form of analog circuitry 80 for performing phase-sensitive detection to extract the signal representing analyte. Unlike many detection problems, the phase and frequency of signal in the present invention are known. Thus, the signal from encoder 72 can be used to perform phase-sensitive detection.

The error signal from the interferometer, previously defined, is at a frequency which is well above the 1/f noise of the amplifiers and so it can be amplified maintaining a good signal-to-noise ratio. To this end, the signal from the interferometer is passed through a bandpass filter (BPF) 82 centered at the signal frequency, and amplified by a low-noise AC amplifier 83.

The amplified error signal is then multiplied with the output from a phase-locked loop (PLL) 84 locked to the reference signal from the encoder by a multiplier 85. PLL 84 is used to convert the encoder signal, typically a square wave, to a sine wave.

As will be shown in detail below, the output of the multiplier has a component at the twice the signal frequency (2F_sig) and a DC component. The DC component is extracted by passing the multiplier output through a low-pass filter (LPF) 87 and the filter output is applied to a DC amplifier 88. The signal has the low noise resulting from low-noise AC amplification and the narrow noise bandwidth resulting from the low-pass filter.

b. Digital Signal Processing

Fig. 6 is a block diagram of circuitry 90 for performing digital data acquisition and analysis to extract the signal information representing bound analyte. In this embodiment, the error signal (or control signal) is digitized by an analog-to digital (A/D) converter 92 and communicated to a computer 95. The A/D converter is clocked at its trigger input by a reference signal from the encoder so that the error signal is sampled many times for each rotation of the disc. Preferably, the sampling rate is greater than the signal frequency so that data can be acquired for each active region at a number of beam positions on that active region.

Computer 95 may have an associated digital signal processor (DSP) for accelerating certain computations such as Fourier transforms. The use of DSPs to take some of the computational load from a main processor is a well-known technique, and will not be discussed further.

Use of a computer to store all of the signal measurements and reference measurements and perform digital averaging has several advantages over the analog technique. No information is lost by doing the signal processing in this fashion. If the data is recorded digitally, each measurement is tagged with information on where the measurement was made on the substrate and the reference signal information is also recorded so that it is then possible to process the data in many different ways.

For example, (1) the data could be analyzed exactly like the analog phase-sensitive detection by multiplying the signal by the reference and then low-pass filtering, (2) the signal could be convolved with a template function (one possible template function would be the measurement of the disc before incubation) , (3) the data could be high-pass filtered and then cut into N strings of data, each of which exactly corresponded to one period of the reference signal and then these strings of data could be averaged point by point, (4) all of the locations on the substrate could be measured several times and then the data at each of these locations could be averaged before the data was further averaged or filtered, (5) all of the measurements could be recorded before signal processing, making it possible to perform a statistical analysis of the individual locations on the substrate to look for bad data caused by imperfections in the substrate or contamination, (6) measurements can be made before incubation and after incubation to look for the buildup of signal corresponding to bound analyte and to subtract away fixed pattern noise from the disc. Indeed, digitally recording the data makes it possible to analyze the data several different ways and compare the results. The measurement locations can be in any pattern on the substrate as long as the computer knows when it is looking at data from an active region and when it is not.

IV. Feedback Control and Servo Amplifier

As mentioned above, the error signal is amplified and bandwidth controlled by servo amplifier 62, which is a low noise, high-gain servo amplifier with gain G_servo(f) that is a function of frequency, f, and then further amplified to high voltage by high-voltage amplifier 65, with gain G_HV. The output voltage V_control of the high-voltage amplifier is called the control signal and is given by:

^Vcontrol ^{= G}HV^Gservo(^f)^Grec^RT^RP0^cos(^Δ) When the servo amplifier is on, the control voltage applied to the PZT will move mirror 35 until the error signal is zero.

As discussed above, the operating point of the interferometer is such that

Δ = π/2 + nπ (where n is an integer) Disturbances from this operating point that are within the bandwidth of the servo will be suppressed by the feedback loop.

If δ denotes the small disturbances to the phase difference, then the control voltage is approximately given by ^vcontrol ~ ^GHV^Gservo^^f)^Grec^RT^RP0^δ Within the servo bandwidth, the control signal is a negative image of the disturbance to the interferometer.

Since any phase difference Δ that is π/2 radians or differs from π/2 radians by an integral multiple of π radians will provide maximum sensitivity (since the absolute value of the derivative of signal power with respect to Δ is maximum) , any of these values represents a suitable operating point for the interferometer. It should be noted, however, that a given servo system will only lock to (i.e., be stable for) points where the slope is positive or to points where the slope is negative. Which slope will be selected depends on the connections of the photoreceiver outputs to the differential amplifier and the polarity of the connection to the PZT. Thus, for one set of connections, the servo system will lock to phase differences of the form π/2 + nπ where n is an even integer (i.e., ... -3π/2, π/2, 5π/2, ...), while for another set of connections, it will lock to phase differences of the form π/2 + nπ where n is an odd integer (i.e., ... -7T/2, 3π/2, 7π/2, ...) . It generally does not matter which way the system is set up since the PZT will generally have sufficient range (say lOπ radians or more) to allow a stable point to be found. If there is a reason that it is desired to accommodate both possibilities, provision can be made to switch the connections. This can be done, for example, by providing a crossover switch at the inputs to differential amplifier 60 or providing an inverter and a unity gain amplifier in a switched arrangement so one of the two can be interposed in the controfl loop (e.g., at the output of servo amplifier 62) . Next considered are feedback schemes that can be used for controlling mirror 35 to keep the interferometer at the operating point. These feedback arrangements include: a) narrowband, b) broadband and c) split feedback.

As mentioned above, the signal representing bound analyte can be measured from either the error signal or the control signal, depending on the servo bandwidth. Since the error signal and the control signal are proportional, at the signal frequency the signal-to-noise ratios of the two are equal. However, there are several practical differences between the two as will be discussed below.

a. Narrowband Servo Excludes the Analyte Signal If the servo bandwidth is less than the signal frequency f_sig then the disturbances in phase produced by the analyte cannot be suppressed by the servo system. Thus the photodetector output (and the error signal) will contain a significant component at the signal frequency, and the control signal will have a reduced component at the signal frequency. The control signal will contain information on the environmental disturbances to the interferometer, for example temperature changes and vibrations, and variations in the thickness of the disc between the two points of impingement of the signal and the reference beams. The temperature variations will be at very low frequency provided the interferometer is thermally insulated from the environment and the thermal mass of the interferometer is large. Vibrations can be attenuated by using a simple seismic isolation stack consisting of alternating layers of metal (e.g., lead, steel, iron) and rubber to support the interferometer.

The absolute variations in the disc thickness are not important. What is important is only the difference in the thickness at the positions of the reference and the signal beams. This is the factor that changes the phase measured by the interferometer. Since the distance between the two beams is kept to a few millimeters or less, provided that the disc thickness does not change rapidly over this small distance, the phase disturbances will be small and at low frequency. The frequency at which these disturbances occur will be near the rotation frequency Ω of the disc and not the much higher signal frequency f_sig.

With a narrowband servo, the servo amplifier filters out most of the high-frequency analyte signal and provides a low-bandwidth error signal that contains all of the large- phase-shift, low-frequency disturbances that would move the interferometer away from the operating point. The analyte signal is then taken at point A between the differential amplifier and the servo amplifier in Fig. 1. This high- frequency signal can then be amplified with a narrowband low- noise AC amplifier and then measured using phase-sensitive detection as described above.

In this case the servo suppresses the low-frequency disturbances and so they cannot appear at the error point. Thus there is less chance of saturating the amplifiers or exceeding the dynamic range of the A/D converter. Also it is easier to implement a narrow-bandwidth servo because the large-dynamic-range PZT's that are typically used have a narrow bandwidth. The random electronic and environmental noise is at frequencies of a few kilohertz or less and so it is preferred for the analyte signal to be at a few kilohertz or above. This is usually easier to accomplish if the analyte signal is on the error signal since the error signal frequency is not limited by the servo bandwidth, provided that the photoreceivers and differential amplifier have sufficient bandwidth.

b. Broadband Servo Includes the Analyte Signal If the servo bandwidth is greater than the signal frequency f_Si_g then the disturbances in phase produced by the analyte are more effectively suppressed by the servo amplifier. Thus the photodetector output will contain only a small component at the signal frequency, but the control signal at point B in Fig. 1 will contain a significant component at the signal frequency. In this case, the analyte signal is measured from point B.

When the servo amplifier has high gain at the signal frequency, the control signal becomes almost independent of the various amplifier gains and depends only on the PZT actuator constant. This could be an advantage if drifts in amplifier gains were a problem. However, the low-frequency variations of the control signal due to disc thickness variations and environmental disturbances now appear at the control point and these large amplitude variations could saturate the low-noise amplifiers or exceed the dynamic range of an A/D converter.

c. Split Feedback Schemes

If PZT 52 does not have sufficient range of motion due to the magnitude of low-frequency disturbances, a second actuator (e.g., another PZT or a voice coil) can be placed in the interferometer. This second actuator is chosen to have large range of motion but low bandwidth. A simple approach is to communicate the control signal to both actuators. Each actuator will respond to the frequency content of the control signal that is within that actuator's bandwidth. This tends to work well, so long as both actuators can withstand the voltage that characterizes the control signal. If this is not the case, separate high-voltage amplifiers, and possibly also separate servo amplifiers can be provided. This amounts to providing two separate control signals. One contains the high- frequency information and is fed back to the high-frequency actuator; the other contains the low-frequency information and is fed back to the low-frequency actuator. In this way both broad bandwidth and large displacement are obtained. The second actuator could be interposed between mirror 37 and mirror mount 50 (Fig. 1) .

V. Disc Configuration and Fabrication

Substrate 20 may be any solid phase material which is at least partially transparent to permit passage of the interferometer beams and which is suitable for covalent or non- covalent (e.g., passive adsorption or ionic interaction) attachment of anti-analyte and other non-specific binding substances. Particularly suitable materials include plastics, such as acrylics, and glass. The substrate may have any geometry which permits it to be mounted in interferometer subsystem 12 and, preferably, to be rotated, translated, or otherwise moved relative to the light beams of the interferometer subsystem. Most preferably, the substrate will be a thin disc formed from acrylic, typically having a thickness in the range of 0.3-3 mm and a diameter in the range of 25-50 mm.

The anti-analyte, usually antibody, antigen, ligand, or antiligand, will be immobilized on at least one surface of the disc, preferably in a regular pattern, i.e., a pattern where the active regions defined by the anti-analyte will pass successively at fixed time intervals through the interferometer light beam as the disc is rotated at a constant speed.

As discussed above, in some embodiments, the disc also has a non-specific binding substance immobilized at regular intervals thereon to define inactive regions. The non- specific binding substance does not specifically bind analyte but non-specifically binds analyte and substances other than analyte from the sample in amounts similar to the specific binding site. The inactive regions can also be defined by bare substrate, i.e., surfaces with no binding substances bound to the substrate. The active and inactive regions can be formed in virtually any shape, but will usually be spots, wedges, or other regular shapes which are easy to form on the substrate surface. If binding to the active region is measured relative to binding to the inactive region, the difference in optical path will, be due to the specific binding of the target analyte, with the amount of non-specific binding in the inactive region generally offsetting that in the active region in question. For example, if the member of the anti-analyte attached to the active region is a monoclonal antibody specific for the analyte of interest, then a slight variant of this antibody that does not specifically bind the analyte of interest (nor specifically bind any other analyte found in the medium the test is carried out in) , would be attached to the disc at a separate site defined as the inactive region. Both regions on the disc would equally and non-specifically bind substances (other than the analyte) from the sample, e.g., non¬ specific protein binding, where the active region would additionally bind the target analyte of interest. The difference in optical path would thus result solely from binding of the target analyte to the active region.

Another method for making the areas of active and inactive binding agents on a surface comprises a first step of uniformly adhering an anti-analyte, e.g. , antibody, to the disc surface. This followed by a step of selectively deactivating spots, wedges, or the like, of the anti-analyte to yield areas of active and deactivated anti-analyte, e.g., by exposing selected areas selected to a deactivating amount of chemical or UV light.

The anti-analyte bound to the disc is selected to bind with the analyte to be determined in the assay. It can be any member of the binding pairs described above. For example, it can be an antibody; antibody fragment selected from the group consisting of Fab, Fab', or F(ab')₂ fragments; hybrid antibody; antigen; hapten; protein A; protein G; lectin; biotin; avidin; chelating agent; enzyme; enzyme inhibitor; protein receptor; nucleotide hybridizing agent; or a bacteria, virus, Mycoplasmatales, spore, parasite, yeast, or fragment thereof; or combinations thereof.

Exemplary disc 20 of Fig. 2 is preferably a plastic disc on which are coated spots 20a of antibody to antigen A and spots 20b of antibody to antigen B. When the disc is incubated with antigen A there is binding between antibodies to A and antigen A but there is no binding between antibodies to B and antigen A. The spots of antibody to antigen B will thus define inactive regions when disc 20 is used for detection of antigen A. Conversely, the spots of antibody to antigen A would define inactive regions in assays for antigen B provided there is no antigen B in the sample being tested for antigen A. Antibodies A and B, however, will also be able to non-specifically bind substances other than antigens A and B, so it will be expected that any differences in the amounts or thicknesses of substances bound to the antibody regions will result from the amount of target antigen (A or B) in the sample. It is this difference which is to be measured and related back to the amount of target antigen in the sample. Thus, disc 20 has spots 20a and 20b of antibody which are alternately specific for antigen A and not specific for antigen A, respectively. These spots 20a and 20b form a ring around the edge of the disc through which the signal and reference beams will travel. Before the disc is incubated with a sample of antigen A, the phase shift experienced by a beam passing through the spots is the same for all spots. To test a sample for antigen A the sample must be incubated on the sample disc.

During incubation, the antibodies of the active spots 20a will specifically bind antigen A and non-specifically bind some amount of other substances. The antibodies of inactive spots 20b, in contrast, will not bind antigen A except non- specifically, but will non-specifically bind other substances in an amount generally equal to that of spots 20a. Thus after incubation, binding of analyte to spots 20a will result in spots 20a having a greater thickness than spots 20b.

As the disc spins, the signal beam experiences a periodically varying phase shift at f_{S g} = NΩ where Ω is the rotation frequency of the disc and N is the number of active spots 20a on the disc circumference. Similarly the reference beam, which passes through a spot adjacent to that traversed by the signal beam, also experiences the same periodically varying phase shift, but 180° out of phase. VI. Sample Application to Substrate

The substrate may be exposed to the liquid sample by any contacting technique which can provide for specific binding between the analyte and the anti-analyte immobilized on the disc surface. Conventional exposure steps include immersion, pipetting, spraying, spin coating, and the like.

The presently preferred contacting technique is manually dispensing a volume of sample with a pipette. The sample may also be pre-diluted with a diluent buffer prior to contacting to the disc. Typical sample volumes are in the range of 25 - 500 μL, preferably 75 - 100 μL.

VII. Practical Detection Problems a. Disc Wobble The disc will have a tendency to wobble at the frequency of rotation and this will produce an apparent signal at the rotation frequency. This will be below the real signal frequency by a factor equal to the number of active spots or wedges on the disc. The disc, assumed to have thickness T, is inserted at an angle θ , as shown in Fig. 7, and the angle of propagation inside the disc is equal to φ, which is given by Snell's law. sin(0) = n sin( ) Solving for the angle φ we find

φ = arcsin sinθ n

If the distance the light travels in the disc is D, the optical path length 0_disc in the disc when the disc is in at angle θ is

°^di- ⁼ ^ ^{= n} w

However, the optical path outside of the disc also changes because as the disc incident angle changes the beam exit point on the disc moves. Let D* be the separation of the entry and exit points on the disc, measured along the beam direction. The optical path outside of the disc relative to its value for 5=0 becomes O_air = T - __•' = T - Dsin |-θ| .φ = T-

= T - [cos (θ) cos (φ) + sin(θ) sin(φ)] cos (φ)

= T [1 - cos (θ) - sin (θ) tan (φ) ]

O_to_tai = O_{disc +}O_ail = T |l_{+ cos} n_(φ) -cos(θ) - tan(φ) sin(θ)

For small values of θ and φ this becomes, to second order:

°total ~ ^T 1 +

+ n (1 + ■ ) - (1 - - -) - θφ

nφ² θ² .« 2 2 ^Ψ

For small values of θ and φ, Snell's law provides 6=nφ or φ=./n. Substituting θ/n for φ gives:

X θz—²

'total n + + - θ__²-

2n 2 n

[_{n +} (n-l)θ²] 2n

For small angular displacements of the disc the optical path length changes by

ΘOtotal ^(n_1).θ3θ n

For example, if the disc is 1 mm thick within 1° of normal and the excursions are less than 1°, the optical path changes by about 0.1 μm, which is a fraction of the wavelength (0.633 μm) for a helium-neon laser. This is easily within the dynamic range and bandwidth of the servo. A further potential problem if the disc wobbles as it rotates is the angle of incidence will vary. Since the reflection from the disc changes with incident angle, the laser power transmitted through the disc will vary, resulting in power modulation at the photodetector. By selecting the angle of incidence and laser polarization so that the laser beam is at Brewster's angle, the reflections from the disc can be eliminated, and to first order, the power variations at the photodetector due to reflection will be minimized. However, these power variations tend to be at the rotational frequency and so do not interfere with the measurement of the analyte signal.

b. Disc Nonuniformitv Non-uniformity in the disc thickness or refractive index will produce fluctuations in the phase difference between the two beams, and could move the interferometer away from the optimum operating point. These types of non-uniformity tend to occur over distances that are large compared to the spot separation. Therefore, their effect can be reduced by configuring the interferometer so that the two beams are close together where they impinge on the disc. Moreover, these fluctuations in the phase difference will tend to be at a frequency on the order of the disc rotational frequency Ω, which is smaller than the signal frequency f_sig by a factor of N, the number of active regions around the disc.

If the servo is narrowband, these disturbances will be nulled by the servo, whereupon the error signal (point A in Fig. 1) will contain the component at _Si_g that measures the bound analyte but will be substantially devoid of these lower- frequency components. If, on the other hand, the servo is broadband so that the output signal must be derived from the control signal (point B in Fig. 1) , both components will be present in the control signal. However, a signal representative of the component at f_sig can be obtained by applying a bandpass filter to the control signal, or preferably by using phase-sensitive detection or spectral analysis as described above. The disc nonuniformities, such as surface roughness, that occur at the spatial frequency of the spots will appear as noise that cannot be suppressed by the servo. However, these disturbances can be measured before incubation and subtracted from the signal measured after incubation.

A second way to reduce the noise due to the disc thickness variations is to make measurements at as many different positions on the substrate as possible. A disc with the active spots disposed on a spiral path, to be described below, makes this possible. Measuring at the same positions on the substrate over and over again will reduce the random noise, but the same substrate thickness variations will be measured repeatedly and these will not be averaged. The full power of the phase-sensitive detection requires that the measurements be independent and the noise random. The signal processing can then be done either using the phase-sensitive measurements in real time or by digital data processing after the accumulation of signal and reference in the computer.

c. Disc Wedge

If the disc is wedged there is a change in the disc thickness of ΔT over the diameter d of the disc. For a distance between the laser beams a on a disc having a refractive index n, a phase excursion of φ is observed:

φ = 2π (n-1) ^(A.^.^)a λd

For a 1-cm diameter disc with refractive index 1.46 with laser spots 1 mm apart and with a wedge of 1 mrad at 633 nm the phase excursion is about 6 radians. This disturbance is within the servo bandwidth because it occurs at the rotation frequency of the disc and the PZT has more than lOπ radians of throw.

d. Variation in Binding Sites

One source of noise in the measurements results from variations in the amount of anti-analyte bound to the substrate in preparation for the incubation of corresponding analyte.

There are two sources of noise. First, to the extent that the surface concentration varies, the phase shift will vary from spot to spot. Second, in addition the amount of bound analyte will' also vary because the amount of it bound to the surface will vary with the surface concentration of anti-analyte immobilized on the surface in preparation for analyte measurement in the first place. The analyte signal will vary from measurement location to measurement location on the substrate. This variation can be reduced by averaging over many independent measurement locations on the substrate. In addition, if the anti-analyte can be measured before the incubation, the measurement of each bound spot after incubation can be corrected for the variation in anti-analyte.

VIII. Other Interferometer Arrangements and Enhancements a. Lenses to Bring Beams Closer

Fig. 8 shows a variation of the optical train in the interferometer, denoted 12', that brings the beams closer together. This is accomplished by placing a pair of lenses 100 and 102 before and after the substrate. Lens 100 causes the beams to cross each other for passage through the substrate while lens 102 directs them on parallel paths for recombining with the second mirror and beamsplitter. The substrate is placed at a location that provides the desired beam separation relative to the pattern on the disc. Since the beam separation now depends on the longitudinal position of the substrate in the beams, this arrangement accommodates different substrate patterns by translation of substrate position. In addition the light beams are reduced in size so that they can examine smaller regions on the substrate and so allow as many spots as possible on the substrate.

b. Beamsplitter-Mirror Combination Element Fig. 9 shows a single element 110 that combines a beamsplitter and mirror. This lowers cost and complexity and makes the interferometer less sensitive to vibration. The beamsplitter-mirror combination is produced by providing a glass substrate 112 having an anti-reflection coating 113 and a high-reflection coating 115 provided on a first surface, and a partial-reflection coating 117 (serving as a 50/50 beamsplitter) and an anti-reflection coating 118 provided on a second (opposite) surface. The thicknesses of the coatings are greatly exaggerated in the figure. The anti-reflection, high- reflection and partial-reflection coatings are all designed for the incident angle used in a particular interferometer.

When using the beamsplitter-mirror element, a light beam incident on anti-reflection coating 113 passes through and encounters coating 117. The beam is partially transmitted by coating 117 and partially reflected by coating 117 to high reflection coating 115, where the partially reflected beam is fully reflected by coating 115 and exits through anti- reflection coating 118. Fig. 10 shows an implementation of a Mach-Zehnder interferometer 120 where the beamsplitters and mirrors are implemented as a pair of elements 122 and 123, each constructed in the manner of combined beamsplitter-mirror element 110 of Fig. 10. Element 123 is merely rotated 180° relative to the position shown in Fig. 10.

In this embodiment, because element 110 fixes in place the mirror and beamsplitter, it is not possible to use a PZT to produce the desired phase difference between the signal and reference laser beams. This is instead accomplished using a Brewster plate 125, which can be rotated by a small angle using a galvanometer 127 to change the optical path length of one of the signal and reference beams. Alternatively, an electro-optic or variable birefringent (liquid crystal) element could be interposed in the path of one of the beams. Fig. 11 shows a variation on the combined beamsplitter-mirror element for use in a system like that shown in Fig. 10, but where it is desired to avoid placing an additional element in the beam. The figure shows an element 130, which only functions as a beamsplitter, used in association with a PZT-mounted mirror such as mirror 35, PZT 52, and mirror mount 55 as shown in Fig. 1. Beamsplitter element is configured much as element 110 of Fig. 9, except that the high-reflection coating on the first surface is replaced by an anti-reflection coating. The second surface has the partial-reflection coating on a portion and the anti- reflection coating on another portion, as in element 110. It is noted that element 130 differs from a normal beamsplitter, which would have partial-reflection coating on all of one surface and anti-reflection coating on the other.

c. Interferometer with Four Beamsplitters and Four Mirrors Fig. 12 shows an alternative interferometer 140 suitable for use with the present invention. Like interferometer 12 of Fig. 1, interferometer 140 includes a first section, upstream of the substrate, for providing the two beams, and a second section, downstream of the substrate, for recombining the beams for detection. In interferometer 12, each of the first and second sections is implemented as a single beamsplitter and a single mirror. In interferometer 140, each of the first and second sections is implemented as two beamsplitters and two mirrors. The first section of interferometer 140 includes beamsplitters 142a and 142b, and mirrors 145a and 145b. The laser beam is split by beamsplitter 142a and the component reflected by beamsplitter 142a is reflected by mirror 145a and encounters beamsplitter 142b. The portion of this component that is transmitted through beamsplitter 142b defines the first or reference beam, and is shown as a normal-weight solid line. The portion of this component that is reflected by beamsplitter 142b is lost, and is shown as a dashed line.

The component transmitted through beamsplitter 142a is reflected by mirror 145b and encounters beamsplitter 142b.

The portion of this component that is reflected by beamsplitter 142b defines the second or signal beam, and is shown as a heavy-weight solid line. The portion of this component that is transmitted through beamsplitter 142b is lost, and is shown as a dashed line.

The second section of interferometer 140 includes beamsplitters 147a and 147b, and mirrors 150a and 150b. Each of the signal and reference beams is split by beamsplitter 147a. The reflected components of the two beams are reflected by mirror 150a and directed to beamsplitter 147b. The transmitted components of the two beams are reflected by mirror 150b and directed to beamsplitter 147b. Of the four beam components encountering beamsplitter

147b, only the transmitted component of the reference beam and the reflected component of the signal beam are superimposed at beamsplitter 147b so as to interfere. The remaining two components (shown in dashed lines) encounter beamsplitter 147b at different locations and are divided individually by beamsplitter 147b and are directed to the photoreceivers without interference. These are shown in dashed lines as well. Thus each photoreceiver receives three beam components, only one of which (shown as a normal-weight solid line) has had its phase modulation converted to amplitude modulation by beamsplitter 147b.

While interferometer 140. suffers the disadvantage relative to interferometer 12 of having more components and losing much of the light, there is a significant advantage. In particular, beamsplitters 142a and 142b and mirrors 145a and 145b are all oriented at 45°. However, their positions can be easily adjusted to provide a separation between the reference and signal beams that is as small as desired without the need for any beam or beam component to pass close to a sharp edge. Thus diffraction effects are avoided.

d. Interferometer with Two Beamsplitters and Six Mirrors

Fig. 13 shows an alternative interferometer 160 suitable for use with the present invention. In interferometer 160, each of the first and second sections is implemented as a single beamsplitter and three mirrors. The first section includes a beamsplitter 162, and mirrors 165a, 165b, and 167. The laser beam is split by beamsplitter 162, with the reflected component becoming the reference beam and the transmitted component becoming the signal beam.

The reference beam is reflected by mirror 165a and directed to the substrate. The signal beam is reflected by mirrors 165b and 167 and directed to the substrate. Beamsplitter 162 and mirrors 165a and 165b are all oriented at 45°, while mirror 167 is oriented at a slightly different angle. Thus, as in the case of interferometer 12* in Fig. 10, the reference and signal beams are not parallel. As mentioned above, this provides extra flexibility in that the beam separation at the substrate can be adjusted merely by adjusting the substrate position along the beam direction.

The second section of interferometer 160 includes mirrors 170, 172a, and 172b, and a beamsplitter 175, arranged complementarily with respect to the first section. The reference beam is reflected by mirror 172b and directed to beamsplitter 175, while the signal beam is reflected by mirrors 170 and 172a and directed to beamsplitter 175. The signal and reference beams are combined at beamsplitter 175 as in the case of interferometer 12 in Fig. 1.

e. Michelson Interferometer in Reflection

Fig. 14A shows an apparatus 180 including a Michelson interferometer suitable for use with the present invention. The figure also illustrates an embodiment of the invention where the substrate 20 is provided with a reflective coating 182 prior to deposition of the active coating and inactive coating (if any) . The Michelson interferometer comprises a beamsplitter 183 and a mirror 185, mounted to a PZT 187, which is mounted to a fixed mirror mount 190. The substrate's reflective coating defines the other mirrors in the interferometer.

In this embodiment, the laser beam is split into two beams by beamsplitter 183. The reflected beam is directed to the substrate, where it passes through any bound material and coating in its path, is reflected by reflective coating 182, passes again through the bound material and coating in reverse order, and is directed to beamsplitter 183. The transmitted beam is reflected from mirror 185 and directed to the substrate, where it passes through any bound material and coating in its path, is reflected by reflective coating 182, passes again through the bound material and coating in reverse order, and is directed to beamsplitter 183. The two beams, after having passed twice through material in their paths, are combined at the beamsplitter and provide two recombined beams. The transmitted recombined beam is detected while the reflected recombined beam is directed into the laser and lost.

This embodiment is provided to illustrate a number of points, although as will be seen, this embodiment is not presently preferred. Due to the fact that this embodiment operates in reflection, the system is much more sensitive to misalignment since minor angular discrepancies direct the beam away from its original direction. Additionally, while the beams pass through the bound analyte twice, potentially providing enhanced signal, the fact that one of the recombined beams is sent into the laser tends to result in additional noise.

Due to the Michelson geometry with normal incidence on the substrate, the system can only accommodate a single photoreceiver (42) . The photoreceiver signal is communicated to one input of differential amplifier 60. A reference voltage corresponding to half the optical power incident on beamsplitter 183 is communicated to the other input. The reference voltage may be derived from an auxiliary detector (not shown) . The output from differential amplifier 60 is the error signal, which can be used with a servo system (not shown) as described above. As discussed above, it is highly desirable to operate where the optical power on photoreceiver 42 is one- half the total optical power.

Fig. 14B is a simplified schematic of a Michelson interferometer embodiment 190 where both recombined beams are detected. As in the embodiment of Fig. 14A, substrate 20 has a reflective coating, here denoted 192, and the interferometer includes a beamsplitter 193 and a mirror 195. This embodiment differs from that of Fig. 14A in that the beams impinge on the reflectively coated substrate at other than a right angle. This results in the beams being recombined at a different location on beamsplitter than the location where the input beam is divided. It is also possible to operate a Mach-Zender interferometer such as that in Fig. 1 in reflection. In such an embodiment, the substrate, reflectively coated as discussed above, would be substituted for mirror 37 in the optical train, and only one beam would encounter the substrate.

f. Other Types of Interferometers

The interferometers described above are not the only interferometer that can be used in an immunosensor. There are many other types of interferometers that may be used, including: (1) Jamin, (2) Fizeau, (3) Sirks-Pringsheim, and (4) Nomarski. Of these, the Nomarski is particularly interesting because it is used in reflection, can be very small and has the two beam positions very close together. Because the Nomarski interferometer is so small it is easy to conceive of an interferometer that is used like a compact disc reading head which scans the surface of the substrate to make its readings.

IX. Other Disc Geometries and Beam Arrangements a. Spiral Track

Fig. 15 shows a disc 20' with active spots 20a¹ and inactive spots 20b' alternating along a spiral path. This arrangement allows measurements to be taken at many more positions on the disc than would be possible with a single track as shown in Fig. 2. This provides better averaging of the disc imperfections. The initial beam positions are shown as X's with one beam on a spot and the other beam between tracks on the spiral. As the disc spins the measurements are made with the interferometer along the spiral from one spot to the next. One beam passes alternately through active and inactive spots while the other always passes through bare substrate. As the disc spins the measurement points on the substrate will have to be moved toward the center of the disc. This can be done by translating the disc center of rotation or translating the interferometer with respect to the disc center of rotation. If the spots on the disc are small it may be necessary to track the measurement position on the disc by imprinting a set of guiding marks on the disc to indicate the position of a given measurement. These marks could be a set of depressions or bumps imprinted on the disc of known depth at a fixed distance from the spot positions. In this way the light beam positions on the substrate could be known at all times.

This arrangement shows the spots at equal angular intervals and thus decreasing spacing toward the center. This allows measurements to be taken at a uniform rotation rate. If the spots were equally spaced along the spiral path, the signal frequency would shift unless the disc rotation rate was also shifted to compensate for this.

b. Wedge-Shaped Active and Inactive Regions Fig. 16 shows a disc 20" with wedge-shaped active regions 20a" and inactive regions 20b" as shown in Fig. 3 of U.S. Patent 4,537,861, the teachings of which are incorporated by reference herein. This configuration is well-adapted for sampling over a large portion of the disc area. The reference signal needs only to encode the information on the angular position and no tracking is required if the beams are slowly moved toward the center of the disc as the disc rotates. This arrangement has the additional advantage that it is easy to maintain a 50% duty factor between active and inactive regions as the radial position of the beams changes.

c. Different Circular Spot Patterns

Figs. 17A-17H show schematically some of the many combinations of disc and beam configuration that are possible according to the present invention. For simplicity, the discs are all shown as having four active regions (spots) at 90° intervals with the four inactive regions interspersed. As noted above, the actual number of active (and inactive) regions is preferably significantly larger. The discs of Figs. 17A-17C and 17G have the interspersed inactive regions coated with a non-specific binding substance. The discs of Figs. 17D-17F and 17H have uncoated inactive regions interspersed with the active regions. Each of the discs of Figs. 17G and 17H has a large central region, inboard of the active regions, that is coated with a non-specific binding substance.

Each of Figs. 17A-17H shows the beam interaction with the disc in a given orientation with an active region at the top, and the beam interaction after the disc has rotated by 45° (or more generally by an interval that is half the spacing of the active regions). The pairs of X's on the figures represent the first and second (signal and reference) beams.

Figs. 18A-18H show idealized plots of the phase difference as a function of time (or disc angular position) as the disc is rotated for the geometries of Figs. 17A-17H, respectively. The geometry of Fig. 17A and the accompanying plot of Fig. 18A correspond to the geometry of Fig. 2 and the accompanying plot of Fig. 3, but are repeated for ease of reference.

Each plot shows the phase differences before and after incubation. The total signal includes a hatched portion representing the phase difference contribution due to- analyte binding added to an unhatched portion representing the phase difference before incubation. The phase differences before incubation result solely from the active and the non-specific binding coatings on the disc. The vertical dashed lines in each of the plots of Figs. 18A-18H represent positions where the signal beam is centered on an uppermost active region of the disc, as shown on the left side of each of Figs. 17A-17H. Fig. 17A shows a configuration where the first beam passes through the active spot while the second beam passes through the inactive spot. After the disc has rotated by 45°, the first beam passes through the inactive spot while the second beam passes through the active spot. As can be seen in Fig. 18A, and as discussed above in connection with Fig. 3, the phase differences before incubation are zero due to the fact that at any given time both beams are either passing through coated spots, passing through uncoated spots, or undergoing a transition on or off coated spots.

Fig. 17B shows a configuration where the first beam passes through the active spot while the second beam passes through an uncoated portion of the disc. After the disc has rotated by 45°, the first beam passes through an inactive spot while the second beam passes through an uncoated portion of the disci. As can be seen in Fig. 18B, the phase differences before incubation are all of the same sign and occur during those intervals where the first beam is on one of the spots (active or inactive) . However the contributions to the signal after incubation only occur when the first beam is on one of the active spots.

Fig. 17C shows a configuration where the first beam passes through an active spot and the second beam does not pass through the disc. After the disc has rotated by 45°, the first beam passes through an inactive spot and the second beam does not pass through the disc. As can be seen in Fig. 18C, the phase differences are the same as those in the case of Fig. 18B since the constant phase difference due to the substrate thickness does not add to the signal.

Fig. 17D shows a configuration where the first beam passes through an active spot while the second beam passes through an uncoated portion of the disc. After the disc has rotated by 45°, the first beam passes through an uncoated portion of the disc while the second beam passes through the active spot. As can be seen in Fig. 18D, the phase differences before incubation alternate in sign as both beams alternate between being on a spot and being on bare substrate. For the same reason, the contributions due to analyte binding alternate in sign.

Fig. 17E shows a configuration where the first beam passes through an active spot while the second beam passes through an uncoated portion of the disc. After the disc has rotated by 45°, the first and second beams pass through uncoated portions of the disc. As can be seen in Fig. 18E, the phase differences before incubation are all of the same sign and occur during those intervals where the first beam is on one of the active spots. Fig. 17F shows a configuration where the first beam passes through an active spot while the second beam does not pass through the disc. After the disc has rotated by 45°, the first beam passes through an uncoated portion of the disc while the second beam does not pass through the disc. As can be seen in Fig. 18F, the phase differences are the same as those in the case of Fig. 18E since the constant phase difference due to the substrate thickness does not add to the signal. Fig. 17G shows a configuration where the first beam passes through an active spot while the second beam passes through the inboard inactive coated region. After the disc has rotated by 45°, the first beam passes through an inactive spot while the second beam passes through the inboard inactive coated region. As can be seen in Fig. 18G, the phase differences before incubation are offset relative to those in the case of Fig. 18B since the second beam is always on a coated region.

Fig. 17H shows a configuration where the first beam passes through an active spot while the second beam passes through the inboard inactive coated region. After the disc has rotated by 45°, the first beam passes through an uncoated portion of the disc while the second beam passes through the inboard inactive coated region agent. As can be seen in Fig. 18H, the phase differences before incubation are offset relative to those in the case of Fig. 18E since the second beam is always on a coated region.

X. Detailed Derivations a. Optical Power as a Function of Path Difference

This is a detailed derivation of the equation, mentioned above in the Photoreceiver and Servo Overview section, for the optical power at photoreceivers 42 and 45 (Fig. 1) . The electric field at photoreceiver 42 is equal to the coherent superposition of the fields from the two paths through the interferometer. The laser beam output power is P₀ and the electric field at first beamsplitter 32 is E₀. The field that passes through beamsplitter 32 (and travels the signal path) is

J!_> 2 The reason for dividing by V_ rather than simply dividing by 2 is that the beamsplitter divides the power in half and the power is proportional to the square of the absolute value of the electric field. The field that is reflected from beamsplitter 32 (and travels the reference path) is

A

The i is included because there is a 90° phase difference between the reflected and the transmitted beams.

These two beams are recombined on second beamsplitter 40. The first field becomes after transmission through beamsplitter 40

2 ^β

where φ₁ is the total phase accumulated in the signal path. The second field becomes after reflection from beamsplitter 40

E, 0 _ai _

where φ₂ is the total phase accumulated in the reference path.

The electric field E₄₂ at photoreceiver 42 is then the sum of these two fields and similarly for the electric field E₄₅ at photoreceiver 45. These are given by:

E₄₂ = - [e ^iφ- e ^iφ<]

The reason for the sign change between the field on photoreceiver 42 and photoreceiver 45 is that the two paths through the interferometer to photoreceiver 45 have each beam reflecting once and transmitting once through a beamsplitter while for photoreceiver 42 one path only transmits and the other path only reflects from a beamsplitter. To compute the optical power on the detectors, the absolute square of the electric fields is integrated over the area of the photoreceivers. If the laser beam is much smaller than the detector area this becomes

P₄₂ = -^[l-cos(Δ)] P₄₅ = -^[H-cos(Δ)]

where Δ is the phase difference between the two paths (i.e., Δ = φ_x-φ₂) . The output of each photoreceiver is a periodic function of the phase difference. If the power on the two detectors is summed the total power is just the power from the laser P₀.

b. Output from Multiplier

Consider first the output of multiplier 85 (Fig. 5) when the signal from amplifier 83 and the reference signal are at the same frequency, namely f_Si_g. In terms of an angular frequency ω, where ω = 2πf_gig, the two signals are proportional to sin(ωt) and sin(ωt+0) where 0 is a fixed phase difference. The product is proportional to:

sin(ωt) sin(ωt+0)

= sin(ωt) [sin(ωt) cos(0) + cos(ωt) sin(0)]

- sin²(ωt) cos(0) + sin(ωt) cos(ωt) sin(0)

If 0=0, i.e., the two signals are in phase, and the product is proportional to:

sin²ωt = [1 - cos(2ωt) ]/2

Thus, the product has a DC component and a component at 2f_sig. The latter can be removed by the low-pass filter.

Consider next the output from the multiplier for unequal frequencies of the input signal and the reference signal. If the input signal has an angular frequency ( H-ω₀) and the reference signal has an angular frequency ω, the two signals are proportional to sin((ω+ω₀)t) (or sin(ωt+ω₀t)) and sin (ωt+0) where 0 is a fixed phase difference. The product is proportional to: sin(ωt+ω₀t) sin(ωt+0)

= [sin(ωt) cos(ω₀t) + cos(ωt) sin(ω₀t)] *

[sin(ωt) cos(0) + cos(ωt) sin(0)] - sin²(ωt) cos(ω₀t) cos(0) + sin(ωt) cosωt [cos( (ω₀t)sin(0) + sin(ω₀t) cos(0)]

+ cos²(ωt) sin(ω₀t) sin(0)

For non-zero values of ω₀, there is no value of 0 for which the product will have a DC component. Thus the contribution from terms at frequencies removed from the signal frequency will be removed by the low-pass filter.

XI. Phase Shift Due to Analyte Binding

Next considered is the optical path length and phase shift. The optical path length, O, is used to compute the accumulated phase shift rather than the physical distance because the wavelength is a function of the refractive index n at each point along the path and so the phase accumulates at different rates along the optical path depending on the refractive index of the medium. Mathematically the optical path length along a path from a to b is the refractive index n(x) integrated over the path from a to b

O = f n (x.) dx.

The phase accumulated after an optical path length O is given by

φ =2_°

Thus the phase accumulates by 2π in an optical path length 0=λ. The phase variation δ is determined from the difference between the optical path length through the active region and the inactive region b °_ac i_ve ⁼ J n (x) dx = n__ubgtrateT_subst-_:ate +

+ n_agT_ag a b

^i_nacti _e ⁼ J n (x) dx = n_subscιateT_subgcrate + n^T^ + n_ai:rT_ag In these equations the T's are the thicknesses and the n's are the refractive indices of the substrate, anti-analyte, and analyte (ab for anti-analyte and ag for analyte; n_air is assumed to be 1) . The changing phase shift is

= 2π ^f (n_ag-l)

Next considered is the refractive index of the bound layers of analyte, anti-analyte, and non-specifically bound substances (referred to collectively as bound layers) . The refractive index of the bound layers is modeled using the Maxwell Garnet formula. This model assumes that the refractive index of all components of the bound layer is the volume weighted average of the refractive indices of the components making up the layer. It is assumed that a fully packed bound layer has a refractive index of n^full _aa_αg = 1.3 and air has a refractive index of nn_aaiirr — - 11..00.. TThhee formula for the refractive index is then just n ■ V • + n ^fullV ⁿag = V, Tot

For low concentrations, the volume of bound layer is small and in a unit area on the disc the formula becomes approximately

n ¹ __g " l-¹- +n"a^fg^u11 _v ^{Vag V}TOt

Now if the individual bound molecules have thickness T, base area w² and surface concentration p, then the volume of all of the components in a bound layer covering a unit area. A, on the disc is

V_ag = Apw²T The total volume that could be occupied is

V_tot = AT Substituting in for the two volumes, n_ag « l + n_a ^f _g ^ullw²p

The phase shift then becomes 2π - _λ ____nr-_a ^f _g ^ullτ_ 2"p

XII. Sensitivity Calculations

The following are computations of the sensitivity of the technique if the source of noise is photon shot noise, for a Mach-Zehnder interferometer.

a. Shot Noise:

To determine the sensitivity, an estimate for the detection noise is needed. Shot noise is assumed for this calculation, which for light of frequency v and power P, in a bandwidth B with detector quantum efficiency η is just

P-^BR²

b. Minimum Detectable Surface Concentration: The minimum measurable signal occurs when the signal equals the noise,

= 1

Substituting in for the signal and solving for the minimum surface concentration,

To determine the sensitivity or detection limit begin by substituting into the equation for the minimum detectable surface concentration the following values. he = 6 . 6xl0^~34 [Js] 3xl0⁸ [m/s] = 2 . OxlO^'25 [ Jm] η = 0. 8

n -._a £_g 11 ≥ ^ 1₁ . 3 _

P = 10^{" 3} [W] λ = 6. 33xl0 -^"7' ,[m]

T_ag = 10-⁹ [m] _w2 _ ₁₀ - ¹⁸ [m² ]

B = 1 [HZ ]

From these numbers is obtained a minimum detectable surface concentration in dimensionless form

λ lmW IOA^' Γ IOA] B

Pmi_n = 1.5X10⁶ 633nnΛ P . T ag . w \ 1Hz

c. Minimum Detectable Solution Concentration

If the total spot area is 1 mm² and the material in the bound layer has a weight of 1000 gm and are in a 0.1 μL sample the solution concentration is

Mass = 1000 l.SxlO⁶ = 2.5xl0^_15[gm] 6. OxlO²³

Concentration = Mass _ 2.5xl0^"15 [gm] _{= 2} 5 pg Volume 0.1μL mL

XIII. Experimental Apparatus

An interferometer configured as shown in Fig. 12 was constructed and used for performing interferometric assays as described below. A helium-neon laser operating at a wavelength of .6328 μm was used.

Electronic apparatus as shown in Fig. 6, with an IBM- compatible personal computer, was used. The A/D converter was triggered by a signal from a Hewlett-Packard HEDS 5540 optical encoder, which provided 512 pulses for every rotation of the motor and disc. The servo amplifier was an LF412 op amp configured as an integrator creating a servo loop unity gain point (bandwidth) of about 500 Hz. The discs had 32 active regions and were rotated at 100 rotations/second, which corresponded to a signal frequency of 3.2 kHz. The signals were taken from the output of the differential amplifier, and passed through a longpass filter having a roll-off of 40 db/decade below 3.2 kHz. A bandpass filter could also be used in the configuration of interest.

XIV. Experimental Materials and Methods a. Disc Spotting Acrylic discs were obtained from Germanow-Simon

(Hesalite HT: crosslinked, uv-stabilized, and heat resistant) . The discs had a diameter of 35 mm, a thickness of 0.6 mm, and a 2.4 mm diameter hole in the center. 100-nL volumes of 40-μg/mL biotinylated bovine serum albumin (biotin-BSA) were dispensed to create 32 spots spaced equidistant at a 1-cm radius. The biotin-BSA volumes were applied with an "Easy Step" dispenser from Tri-Continent Scientific. This was performed by contacting the syringe tip of the dispenser to the acrylic surface. Fluid volumes were dispensed by positive displacement with the syringe plunger. This resulted in spots, each having an area of approximately 0.8 mm², with a spot center-to-center spacing of approximately 2 mm. After the spots were allowed to dry, the discs were washed several times with deionized water. The adsorbed biotin-BSA serves as the immobilized solid phase component in both assay protocols described below.

b. Protocol for Detection of Biotin

1. A 25 μL biotin sample was mixed with 250 μL of a solution containing streptavidin-horseradish peroxidase conjugate (HRP Pierce #21726) and incubated for 10 minutes at room temperature. The buffer formulation for both sample and conjugate was phosphate buffered saline (PBS) with 0.1% Tween^® 20 and 1 mg/mL BSA at pH 7.2.

2. The resulting mixture was placed with a pipette on the biotin-BSA disc and incubated for 1 hour at room temperature, then washed 4 times with PBS, 0.1% Tween^® 20. 3. 300 μL of dia inobenzidine/peroxide substrate (Pierce # 34065) was added to the disc and incubated for 10 minutes at room temperature.

4. The discs were washed with deionized water, followed by measurements in the interferometer system as illustrated in Fig. 12 above.

c. Protocol for Detection of CKMB

1. 400 μL of solution containing anti-CKMB streptavidin conjugate in phosphate buffered saline (PBS), 0.1% Tween^® 20, lmg/mL at pH 7.2 was placed on the disc and incubated for one hour at room temperature. The discs were then washed four times with 400 μL volumes of phosphate buffered saline. The anti-CKMB was a murine monoclonal antibody.

2. 400-μL CKMB samples in phosphate buffered saline pH 7.2 were placed on the disc, incubated for one hour at room temperature, then washed four times with PBS. The CKMB samples ranged from zero to 625 ng/mL in concentration. 3. 400 μL of a solution containing goat anti-CKMM antibody in PBS, 0.1% Tween^® 20, 1 mg/mL BSA at pH 7.2 was placed on the disc, incubated for 30 minutes at room temperature, then washed four times with PBS, 0.1% Tween^® 20.

4. 400 μL of a solution containing rabbit anti-goat IgG peroxidase conjugate was placed on the disc, incubated for

30 minutes at room temperature, then washed four times with PBS.

5. 400 μL of diaminobenzidine/peroxide substrate solution (Pierce #34065) was added to the disc and incubated for 10 minutes at room temperature. The discs were washed in deionized water, then dried.

6. The discs were measured in the interferometer as illustrated in Fig. 12 above.

d. Interferometer Calibration

The interferometer was first calibrated to determine the slope of the response curve at the operating point. To this end, the servo was turned off and a sawtooth at approximately 100 Hz was applied to the PZT in order to cause the phase difference between the two paths to vary over more than¹2π radians. A peak measurement circuit was used to find the peak-to-peak voltage of the error signal. This was also observed on an oscilloscope. This measurement yielded a value of the voltage change occurring when the interferometer was scanned over 2_ radians of phase difference. This allowed a conversion from voltage to thickness in Angstroms to be calculated. A typical value for this calibration constant was approximately 7.5 millivolts/Angstrom. This calibration constant was used to derive the y-axis values on the binding assay plots.

e. Data Acquisition and Analysis Many rotations of the sample discs were digitized and the data values were averaged on a point-by-point basis. The computer then calculated the power spectral density (PSD) using a fast Fourier transform (FFT) routine. The Fourier component at 32 times the motor rotation speed (3.2 kHz in this case) was read out as the signal output of the measurement. Measurements were made before and after incubation, and the results subtracted.

XV. Experimental Results a. Binding Assay Measurements

Fig. 19 is a graph illustrating the thickness of a biotin layer bound to a disc substrate and measured with an interferometer according to the method of the present invention, where the thickness varies inversely with biotin concentration in a competitive assay format. These results demonstrate the ability to bind labelled biotin in a competitive assay format in thicknesses which are readily measured using an interferometer according to the method of the present invention. The fact that the signal changed measurably between a concentration of zero and a concentration of 1 picogram/milliliter suggests that the technique of the present invention is highly sensitive. Fig. 20 is a graph illustrating the thickness of a CKMB-containing layer bound to a disc substrate and measured with an interferometer according to the method of the present invention, where the thickness varies directly with CKMB concentration in a sandwich assay format. These results demonstrate the ability to bind CKMB in a sandwich binding assay format in thicknesses which are readily measured using an interferometer according to the method of the present invention.

b. Interferometer Signal Traces

Oscilloscope measurements were taken at the output of the differential amplifier and at the output of the servo amplifier. The interferometer was of the type shown in Fig. 12. The disc was rotated at a uniform rate of 100 rotations/second, which for 32 active spots corresponds to a signal frequency of 3.2 KHz. The same servo system, having a bandwidth of about 500 Hz, was used for both measurements, and represents the situation where the signal frequency is beyond the servo bandwidth.

Fig. 21 is a plot of the interferometer voltage signal as a function of time, taken with the oscilloscope connected to the differential amplifier output. This plot represents the signal at point A in Fig. 1. Since the signal frequency was outside the bandwidth of the servo, the servo suppressed much of the low-frequency phase disturbances but not the phase disturbances at the signal frequency.

The low-frequency component that remains in the plot (i.e., the portion that was not suppressed by the servo) has frequencies generally in the range of a few times the rotational frequency. The particular nature of this residual signal depends on the servo bandwidth and the spatial frequency distribution of disc non-uniformities and the like.

Fig. 22 is a plot of the interferometer voltage signal as a function of time, taken with an oscilloscope connected to the servo amplifier output. This plot represents the signal at point B in Fig. 1. Since the plot shows a signal proportional to the control signal, the component at the rotational frequency, as well as components at a few times the rotational frequency, appear, along with a reduced amount of signal at the signal frequency. The overall DC level depends on the settling point of the servo.

XVI. Conclusion

In conclusion it can be seen that the present invention provides a sensitive and accurate binding assay technique. While the above is a complete description of specific embodiments of the invention, various modifications and alternatives are possible. Therefore, the above description should not be taken as limiting the scope of the invention which is defined by the claims.

Claims

WHAT IS CLAIMED IS:

1. For use with a substrate having (a) at least one active region having thereon an immobilized substance that specifically binds an analyte and (b) at least one inactive region, an interferometer-based system for measuring analyte bound to said active region, comprising: a light source for outputting a source beam; first means for splitting said source beam into first and second beams directed along first and second light paths, respectively, with at least one of said first and second beams encountering said substrate; means for effecting relative movement between said substrate and said first and second light paths so that at least one of said first and second beams encounters said active and inactive regions and passes through analyte bound to said active region, wherein at least one of said first and second beams undergoes a phase change that depends on the amount of analyte bound to said active region as a result of encountering said active and inactive regions; second means for recombining said first and second beams after said at least one of said first and second beams encounters said active region, said first beam accumulating a first total phase along said first path between said first means and said second means, said second beam accumulating a second total phase along said second path between said first means and said second means; said second means providing at least one recombined beam having optical power that depends on the difference between said first and second total phases; optical detection means, coupled to said combining means, for converting said at least one recombined beam into at least one electrical signal representing the optical power of said at least one recombined beam; and third means, responsive to said at least one electrical signal, for providing an output signal that depends on the amount of analyte bound to said active region based on said phase change undergone by said at least one of said first and second beams encountering said active and inactive regions.

2. The system of claim 1 wherein: said substrate is optically transparent; and said at least one of said first and second beams passes through said substrate;

3. The system of claim 1 wherein: each of said active and inactive regions has a reflective surface; said reflective surface of said active region has said immobilized substance deposited thereon; said reflective surface of said inactive region has no substances deposited thereon; said at least one of said first and second beams, on encountering said active region, passes through said analyte bound to said active region and through said immobilized substance,, is reflected from said substrate, and passes again through said immobilized substance and through said analyte bound to said active region; and said at least one of said first and second beams, on encountering said inactive region, is reflected from said substrate.

4. The system of claim 1 wherein: each of said active and inactive regions has a reflective surface; said reflective surface of said active region has said immobilized substance deposited thereon; said reflective surface of said inactive region has deposited thereon a substance, referred to as the non-specific binding substance, that non-specifically binds substances other than the analyte; said at least one of said first and second beams, on encountering said active region, passes through analyte bound to said active region and through said immobilized substance, is reflected from said substrate, and passes again through said immobilized substance and through said analyte bound to said active region; and said at least one of said first and second beams, on encountering said inactive region, passes through substances bound to said non-specific binding substance and through said non-specific binding substance, is reflected from said substrate, and passes again through said non-specific binding substance and substances bound to said non-specific binding substance.

5. For use with an optically transparent substrate having (a) at least one active region having thereon an immobilized substance that specifically binds an analyte and (b) at least one inactive region, an interferometer-based system for measuring analyte bound to said active region, comprising: a light source for outputting a source beam; first means for splitting said source beam into first and second beams directed along first and second light paths, respectively, with at least one of said first and second beams passing through said substrate; means for effecting relative movement between said substrate and said first and second light paths so that at least one of said first and second beams is directed through said active and inactive regions, wherein at least one of said first and second beams undergoes a phase change that depends on the amount of analyte bound to said active region as a result of passing through said active and inactive regions; second means for recombining said first and second beams after said at least one of said first and second beams passes through said active region, said first beam accumulating a first total phase along said first path between said first means and said second means, said second beam accumulating a second total phase along said second path between said first means and said second means; said second means providing at least one recombined beam having optical power that depends on the difference between said first and second total phases; optical detection means, coupled to said combining means, for converting said at least one recombined beam into at least one electrical signal representing the optical power of said at least one recombined beam; and third means, responsive to said at least one electrical signal, for providing an output signal that depends on the amount of analyte bound to said active region based on said phase change undergone by said at least one of said first and second beam passing through said active and inactive regions.

6. The system of claim 5, and further comprising: control means, coupled to at least one of said first and second light paths and responsive to a control signal, for modifying the optical path length of at least one of said first and second paths; and wherein said third means is further for generating said control signal so as to maintain a desired relationship between said first and second total phases.

7. (Amended) The system of claim 6 wherein said desired relationship is that the derivative of said at least one electrical signal with respect to said difference between said first and second total phases is a maximum.

8. The system of claim 6 wherein: at least one of said first and second means includes a mirror for deflecting light travelling along at least one of said first and second paths; and said control means includes an actuator to which said mirror is mounted.

9. The system of claim 6 wherein: said control means includes an optical element disposed in one of said first and second paths for passage of one of said first and second beams therethrough.

10. The system of claim 6 wherein: said first means comprises a beamsplitter that passes said first beam and reflects said second beam, and a mirror that deflects said second beam along a path generally parallel to said first beam; and said control means comprises a piezoelectric transducer coupled to said mirror and responsive to an electrical signal for moving said mirror.

11. The system of claim 5, and further comprising: means, coupled to said means for effecting relative movement, for providing a reference signal synchronized with said relative movement.

12. The system of claim 5 wherein: said substrate is a disc; and said means for effecting relative movement includes means for rotating said disc.

13. The system of claim 5 wherein said inactive region has an immobilized substance that non-specifically binds substances other than analyte.

14. The system of claim 5 wherein: said second means provides first and second recombined beams; said optical detection means converts said first and second recombined beams into first and second electrical signals; and said third means provides a signal proportional to the difference of said first and second electrical signals.

15. The system of claim 5 wherein: said first beam repeatedly passes through said active and inactive regions on said substrate; and said second beam does not pass through said substrate.

16. The system of claim 5 wherein: said first beam repeatedly passes through said active and inactive regions on said substrate; and said second beam passes through said substrate, but does not pass through said active region.

17. The system of claim 5 wherein: said first beam repeatedly passes through said active and inactive regions on said substrate; and said second beam repeatedly passes through said active and inactive regions on said substrate.

18. The system of claim 5 wherein: said substrate has a plurality of active regions and a plurality of inactive regions; said first beam passes through at least some of said active and inactive regions on said substrate; and said second beam passes through at least some of said active and inactive regions on said substrate.

19. The system of claim 5 wherein: said substrate has a plurality of active regions and a plurality of inactive regions; said first beam passes through all of said active and inactive regions on said substrate; and said second beam passes through all of said active and inactive regions on said substrate.

20. The system of claim 5 wherein: said substrate has a plurality of active regions interspersed with a plurality of inactive regions along a path; and said first and second beams are spaced relative to said active and inactive regions along said path so that said first beam passes through one of said active regions while said second beam passes through an adjacent inactive region, and after said substrate has moved, said second beam passes through one of said active regions while said first beam passes through an adjacent inactive region.

21. For use with an optically transparent substrate having (a) a plurality of active regions that specifically bind an analyte and (b) a plurality of inactive regions that do not bind the analyte, arranged in a predetermined pattern, an interferometer-based binding assay system for measuring analyte bound to said active region, comprising: a light source for outputting a source beam; first means for splitting said source beam into first and second beams directed along first and second light paths, respectively, with at least one of said first and second beams passing through said substrate; means for moving said substrate relative to said first and second light paths so that at least one of said first and second beams passes alternately through at least some of said active and inactive regions and undergoes phase changes as a result of passing through said active and inactive regions; second means for recombining said first and second beams after said first and second beams pass through said substrate, said first beam accumulating a first total phase along said first path between said first means and said second means, said second beam accumulating a second total phase along said second path between said first means and said second means; said second means providing first and second recombined beams having respective optical powers that depend on the difference between said first and second total phases; and photodetector means, coupled to said second means, for converting said first and second recombined beams into respective first and second electrical signals representing the optical power of said first and second recombined beams; control means, coupled to at least one of said first and second light paths and responsive to a control signal, for modifying the optical path length of at least one of said first and second paths relative to the other; third means, responsive to said electrical signals, (a) for generating said control signal so as to maintain a desired relationship between said first and second total phases, and (b) for providing an output signal that depends on the amount of analyte bound to said active regions based on said phase changes.

22. The system of claim 21 wherein: each of said first and second beams passes alternately through at least some of said active and inactive regions with (a) said first beam passing through an active region when said second beam passes through an inactive region, and (b) said second beam passing through an active region when said first beam passes through an inactive region.

23. The system of claim 21, and further comprising: means, coupled to said means for effecting relative movement, for providing a reference signal synchronized with said relative movement.

24. The system of claim 21 wherein: said substrate is a disc; and said means for effecting relative movement includes means for rotating said disc.

25. The system of claim 21 wherein said inactive regions have an immobilized substance that non-specifically binds substances other than analyte.

26. The system of claim 21 wherein: said first beam repeatedly passes through said active and inactive regions on said substrate; and said second beam does not pass through said substrate.

27. The system of claim 21 wherein: said first beam repeatedly passes through said active and inactive regions on said substrate; and said second beam passes through said substrate, but does not pass through said active region.

28. The system of claim 21 wherein said desired relationship is that the derivative of a difference between said first and second electrical signals with respect to the difference between said first and second total phases is a maximum.

29. The system of claim 21, wherein said first means comprises: a first beamsplitter including a first partially transmitting surface for transmitting as said first beam a portion of said source beam and reflecting as said second beam a portion of said source beam; and a first mirror for reflecting one of said first and second beams toward said substrate.

30. The system of claim 29, wherein: said second means comprises a second mirror for reflecting said first beam after said first beam traverses said substrate, and a second beamsplitter including a second partially transmitting surface for transmitting a portion of the first beam reflected by said second mirror and reflecting a portion of said second beam after said second beam passes through said substrate in a first direction to define said first recombined beam, said second partially transmitting surface reflecting a portion of said first beam and transmitting a portion of said second beam in a second direction to define said second recombined beam; and said photodetector means comprises a first photodetector arranged to intercept beams transmitted in said first direction by said second means, and a second photodetector arranged to intercept beams transmitted in said second direction by said second means.

31. The system of claim 30, wherein: said third means comprises a differential amplifier having positive and negative inputs coupled to respective outputs of said first and second photodetectors, and servo amplifier means, having a predetermined gain and frequency response, for amplifying an output of said differential amplifier and producing an amplified output; and said control means comprises piezoelectric transducer means driven by an output of said servo amplifier means and coupled to one of said first and second mirrors for controlling positioning of said one of said first and second mirrors so that light is incident on said first and second photodetectors in amounts that are substantially the same.

32. The system of claim 31, wherein said servo amplifier means comprises a narrowband amplifier and said output signal is derived from an output of said differential amplifier.

33. The system of claim 31, wherein said servo amplifier means comprises a wideband amplifier and said output signal is derived from an output of said wideband amplifier.

34. The system of claim 31, wherein said substrate has said active and inactive regions arranged as interspersed pluralities of spots disposed in a circle about a rotation axis of said substrate.

35. The system of claim 31, wherein said substrate has said active and inactive regions arranged as interspersed pluralities of wedges of said substrate.

36. The system of claims 31, wherein said first means comprises: a first transparent element having opposed first and second surfaces, a first anti-reflection coating and a first high reflection coating provided adjacent each other on said first surface, and a first partially reflecting coating and a second anti-reflection coating provided adjacent each other on said second surface, said first partially reflective coating and said second anti-reflection coating being opposite said first anti-reflection coating and said first high reflection coating, .respectively.

37. The system of claim 36, wherein said second means comprises: a second transparent element having opposed third and fourth surfaces, a third anti-reflection coating and a second partially reflecting coating provided adjacent each other on said third surface, and a second high reflection coating and a fourth anti-reflection coating provided adjacent each other on said fourth surface, said second high reflection coating and said fourth anti-reflection coating being opposite said third anti-reflection coating and said second partially reflecting coating respectively.

38. The system of claim 37, wherein said control means comprises: a Brewster plate inserted in one of said first and second optical paths.

39. The system of claims 31, and further comprising: means for producing a synchronization signal in synchronization with said output signal; and phase-sensitive detection means having as inputs said synchronization signal and said output signal for producing a low pass filtered signal corresponding to analyte bound to said active regions.

40. A method of measuring an analyte on a substrate, comprising: providing a transparent substrate having (a) at least one active region with analyte specifically bound thereto, and (b) at least one inactive region with no significant analyte bound thereto; providing an interferometer having at least two beams travelling along two beam paths; inserting the substrate into at least one of the beam paths; moving the substrate relative to the interferometer so that at least one of the beams passes through the active region and the inactive region; measuring a phase change in the at least one of the beams passing through the active and inactive regions due to analyte binding to the active region; and generating an output signal indicative of the amount of analyte bound to the active region based on the measured phase change.

41. The method of claim 40 wherein the active and inactive regions have substances other than analyte non- specifically bound thereto.

42. The method of claim 40, and further comprising the step of controlling the interferometer by modifying at least one of the beam paths so as to maintain a desired phase difference between the two beam paths.

43. A method of measuring an analyte in a sample, comprising: contacting a transparent substrate having (a) at least one active region having a substance immobilized thereon which specifically binds the analyte and (b) at least one inactive region whereby analyte in the sample will bind to the active region but not to the inactive region; providing an interferometer having at least two beams travelling along two beam paths; inserting the substrate into at least one of the beam paths; moving the substrate relative to the interferometer so that at least one of the beams passes through the active region and the inactive region; measuring a phase change in the at least one beam passing through the active and inactive regions due to analyte binding to the active region; and generating an output signal indicative of the amount of analyte bound to the active region based on the measured phase change.

44. The method of claim 43, and further comprising the step of controlling the interferometer by modifying at least one of the beam paths so as to maintain a desired phase difference between the two beam paths.