US20100122904A1 - Incorporating cmos integrated circuits in the design of affinity-based biosensor systems - Google Patents

Abstract

A biosensor system incorporating CMOS integrated circuits. In one type of biosensor system, the biosensor system includes a complementary metal-oxide-semiconductor (“CMOS”) integrated circuit. The biosensor system further includes an optical filter fabricated on the CMOS integrated circuit. Additionally, a plurality of capturing probes is optically coupled to the CMOS integrated circuit. Alternatively, another type of biosensor system includes a silicon substrate. The alternative biosensor system further includes active devices fabricated on the silicon substrate. Additionally, the alternative biosensor system includes a plurality of metal layers stacked on top of the active devices. Furthermore, the alternative biosensor system includes a passivation layer covering a top metal layer, where the passivation layer includes an opening configured to expose the top metal layer, where the opening is used as a sensing electrode. Additionally, the alternative biosensor system includes a plurality of probes attached to the sensing electrode.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS
• This application is related to the following commonly owned co-pending U.S. patent application:
• Provisional Application Ser. No. 61/115,485, “Incorporating CMOS
• Integrated Circuits in the Design of Affinity-Based Biosensor Systems,” filed Nov. 17, 2008, and claims the benefit of its earlier filing date under 35 U.S.C. §119(e).
TECHNICAL FIELD
• The present invention relates to biosensor systems, and more particularly to incorporating CMOS integrated circuits in the design of affinity-based biosensor systems.
BACKGROUND OF THE INVENTION
• Affinity-based detection is a fundamental method to identify and measure the abundance of biological and biochemical analytes and is one of the most important analytical methods in biotechnology. Affinity-based detectors (or so-called biosensors in case of detecting biological analytes) take advantage of the selective interaction and binding (affinity) of the target analyte with immobilized capturing probes to specifically capture the target analyte onto a solid surface. A goal of a detection platform is to facilitate specific capturing and ultimately to produce a detectable signal based on the captured analytes. The generated signals correlate with the presence of the target analytes in the sample (e.g., toxins, polymers, hormones, DNA strands, proteins, bacteria, etc.), and hence are used to estimate their abundance.
• To create target-specific signals in biosensors, the target analytes in the sample volume first need to collide with the capturing layer, interact and bind to the probes, and ultimately take part in a transduction process (i.e., a physiochemical process which produces certain measurable electrical, mechanical, or optical parameters produced solely by the captured entities). The analyte motion in typical biosensor settings (e.g., aqueous biological mediums) is dominated by diffusion spreading, which from a microscopic point of view is a probabilistic mass-transfer process (i.e., random walk events for a single analyte molecule). Accordingly, the analyte collisions with the probes become probabilistic processes. Moreover, because of the quantum-mechanical nature of chemical bond formation, interactions between probes and analytes, are also probabilistic, adding more uncertainty to the capturing procedure. On top of these two processes which can be considered the biochemical noise of the system, there may also be a detector and a readout circuitry (e.g., optical scanners for fluorescent-based transducers), which likely add additional noise to the already noisy signal.
• Besides the inevitable uncertainty associated with the target analyte capturing and detecting, in all practical biosensors, binding of other species to the probes (non-specific binding) is also possible. Non-specific binding (e.g., cross-hybridization in DNA microarrays) is generally less probable than the specific binding when target analytes and the interfering species have the same abundance. Nonetheless, when the concentration of the non-specific species becomes much higher than the target analyte, non-specific bindings (or essentially interference) may dominate the measured signal and hence limit the minimum-detectable-level (MDL). In biosensors, the MDL may be either biochemical noise or interference-limited, while the highest detection level (HDL), is solely a function of capturing probe density and its saturation level.
• Due to such impediments, as of today, the accuracy of biosensors systems does not satisfy the stringent requirements of many high-performance biotechnology applications in molecular diagnostics and forensics. In addition, biosensors systems have not successfully made the transition to portable and compact point-of-care devices because their detection platforms still consist of fluidic systems and bulky detectors.
• One proposed solution to address the challenges of biosensor systems is to use semiconductor fabrication technologies to build compact, high-performance, and cost-efficient biosensor systems. It is envisioned that such systems (i.e., lab-on-a-chip platforms), include not only the fluidic (macro or micro) systems and sample preparation processes, but also the integrated transducers.
• The challenge of designing sample preparation modules in biosensors, to some extent, has been addressed in recent years, particularly in the form of micro-fluidic and automated liquid handling systems; however, the integration of the detector and readout circuitry has not been addressed. One reason why the integration of the detector and readout circuitry has not been addressed is the technical challenge of manufacturing transducers using custom surface and bulk MEMS procedures. Another reason is performance and cost justification of monolithic integration of all components.
• In recent years, the idea of employing Complementary Metal-oxide-semiconductor (CMOS) fabrication processes, which are the most robust and widely used fabrication processes in the semiconductor industry, for biosensors has emerged. The rationale behind this, as opposed to using MEMS or other custom processes, is the unmatched yield, cost-efficiency, and the integration capabilities of CMOS processes. While CMOS processes, from the electronic design point of view, offer huge degree of design flexibility and system integration, they are not very flexible in terms of form factor, transducer design and interface integration. Challenges remain in designing biosensors to take advantage of the CMOS fabrication method. The primary design challenge using CMOS technology is the interface design between the assay and integrated chip (IC) which requires additional post-fabrication processes for compatibility in detecting targets (e.g., analytes).
• Therefore, there is a need in the art for incorporating the use of CMOS fabrication processes in the design of affinity-based biosensor systems.
BRIEF SUMMARY OF THE INVENTION
• In one embodiment of the present invention, a biosensor system comprises a complementary metal-oxide-semiconductor integrated circuit, where the complementary metal-oxide-semiconductor integrated circuit comprises a silicon substrate and a dielectric layer on the silicon substrate. The biosensor system further comprises an optical filter fabricated on the complementary metal-oxide-semiconductor integrated circuit. Furthermore, the biosensor system comprises a plurality of capturing probes optically coupled to the complementary metal-oxide-semiconductor integrated circuit.
• In another embodiment of the present invention, a biosensor system comprises a silicon substrate. The biosensor system further comprises active devices fabricated on the silicon substrate. Additionally, the biosensor system comprises a plurality of metal layers stacked on top of the active devices. Furthermore, the biosensor system comprises a passivation layer covering a top metal layer of the plurality of metal layers in order to protect the plurality of metal layers, where the passivation layer comprises an opening configured to expose the top metal layer, where the opening is used as a sensing electrode. Additionally, the biosensor system comprises a plurality of probes attached to the sensing electrode.
• The foregoing has outlined rather generally the features and technical advantages of one or more embodiments of the present invention in order that the detailed description of the present invention that follows may be better understood. Additional features and advantages of the present invention will be described hereinafter which may form the subject of the claims of the present invention.
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING
• A better understanding of the present invention can be obtained when the following detailed description is considered in conjunction with the following drawings, in which:
• FIGS. 1A-D illustrate a fully integrated fluorescent-based biosensor microarray system in accordance with an embodiment of the present invention;
• FIG. 2 is a graph of the filter transmittance and photodiode quantum efficiency versus the wavelength in accordance with an embodiment of the present invention;
• FIG. 3 illustrates optically coupling the capturing probes with the CMOS detector using a fiber-optical faceplate in accordance with an embodiment of the present invention;
• FIG. 4 illustrates an embodiment of the present invention of a pixel;
• FIG. 5 is a graph illustrating the measured sensitivity of the integrated microarray and the background current present in the system as a function of temperature in accordance with an embodiment of the present invention;
• FIG. 6 is a graph illustrating the measured real-time DNA hybridization kinetics for three exemplary DNA target concentrations in solution in accordance with an embodiment of the present invention;
• FIG. 7 illustrates impedance spectroscopy-based detection in accordance with an embodiment of the present invention;
• FIGS. 8A-B illustrate a pixel with an on-chip electrode that is formed by creating an opening in a passivation layer covering the metal layers in accordance with an embodiment of the present invention;
• FIG. 9 illustrates a scanning electron microscope (SEM) photograph of a surface of the fabricated pixel ofFIGS. 8A-B in accordance with an embodiment of the present invention;
• FIG. 10 illustrates the excitation scheme of a chip in accordance with an embodiment of the present invention;
• FIG. 11 illustrates the basic concept used in measuring impedance by the detection circuitry in accordance with an embodiment of the present invention; and
• FIG. 12 illustrates a circuit for the each individual pixel in accordance with an embodiment of the present invention.
DETAILED DESCRIPTION OF THE INVENTION
• In the following description, numerous specific details are set forth to provide a thorough understanding of the present invention. However, it will be apparent to those skilled in the art that the present invention may be practiced without such specific details. In other instances, well-known circuits have been shown in block diagram form in order not to obscure the present invention in unnecessary detail. For the most part, details considering timing considerations and the like have been omitted inasmuch as such details are not necessary to obtain a complete understanding of the present invention and are within the skills of persons of ordinary skill in the relevant art.
• As discussed in the Background section, biosensors are one of the most important analytical tools in biotechnology today. These detection systems take advantage of the selective interaction and binding of certain biological molecules to identify and detect different analytes such as toxins, hormones, DNA strands, proteins, bacteria, etc. The fundamental advantage of array based biosensors, which compensate for their limited signal-to-noise ratio (“SNR”), is their capability to detect multiple analytes simultaneously. Today, densely packed biosensor arrays (i.e., microarrays) which detect hundreds or even thousands of different analytes are an integral part of biotechnology.
• Certain emerging biotechnology applications, such as high-throughput molecular screening and point-of-use (PoU) molecular diagnostics, necessitate biosensor integration, particularly the interfacing of the biochemical part (assay) with the transducer and the readout circuitry. This is mainly due to the stringent requirements of applications which demand compact, cost-efficient, and disposable systems with a high production yield and robust functionality; a goal which silicon-based integrated circuits technology in general, and CMOS processes in particular can provide.
• The dominant biosensor and microarray detection modality is visible-range fluorescence spectroscopy using fluorescent labels as the reporters for target analyte molecules. While alternative “label-free” transduction methods (e.g., electrochemical or magnetic) exist today, fluorescent-based detection still remains the most sensitive and robust method, particularly in DNA detection application. The performance advantages of this detection method over other methods originate from the uniqueness of fluorescence phenomenon which makes the generated signals very specific and less susceptible to biological interference.
• Referring toFIGS. 1A-D,FIGS. 1-D illustrate a fully integrated fluorescent-basedbiosensor microarray system100 in accordance with an embodiment of the present invention.System100, as shown inFIG. 1A, includes a packaged CMOS integratedcircuit101 with asensing area102. In one embodiment,sensing area102 has a thickness of approximately 3 millimeters. Details ofsensing area102 are provided further below in connection withFIG. 3 showing the transducers, emission filter, and readout circuitry, including the analog-to-digital converter (ADC).
• To visualize the structures ofsensing area102 for biological significance, afluorescent image103 of a portion ofsensing area102 is taken as shown inFIG. 1B. Fluorescent imaging techniques, include, but are not limited to electron microscopy, x-ray crystallography, nuclear magnetic resonance spectroscopy and atomic force microscopy.Fluorescent image103 illustrates photo-detectors104A-C. Photo-detectors104A-C may collectively or individually be referred to as photo-detectors104 or photo-detector104, respectively.System100 may include any number of photo-detectors104 and the number of photo-detectors104 shown inFIG. 1B is illustrative.
• Fluorescent image103 further illustrates targets (e.g., analytes)105A-B. Targets105A-B may collectively or individually be referred to astargets105 ortarget105, respectively.System100 may include any number oftargets105 and the number oftargets105 shown inFIGS. 1C and 1D is illustrative.
• For eachtarget105, a fluorescent label is captured by a DNA capturing probe. For example, capturingprobes106A-C are used to capture fluorophores107A-B using a process referred to as fluorescent labeling in connection withtarget105A. Similarly, capturingprobes106D-F are used to capturefluorophore107C in connection withtarget105B. Capturing probes106A-F may collectively or individually be referred to as capturingprobes106 or capturingprobe106, respectively. Further, fluorophores107A-C may collectively or individually be referred to asfluorophores107 orfluorophore107, respectively.System100 may include any number of capturingprobes106 andfluorophores107 and the number of capturingprobes106 andfluorophores107 shown inFIGS. 1C and 1D is illustrative.
• Whilesystem100 is designed and fabricated for DNA microarrays, the achieved specifications are well suited for other biosensor applications.
• The foremost challenge in designing fluorescent-based detectors is the proper excitation of labels and the detection of their emitted signal. The photon absorption of the fluorescent label, denoted by A inFIGS. 1A-D, exposed to an incident photon flux, F_X, obeys the Beer-Lambert law. For a thin layer of fluorescent labels, the absorption is given by
• 
  A=F_X[1−e^−a0(λ)N]≈F_Xa₀(λ)N,  (1)
• where N is the surface concentration of labels with extinction coefficient of a₀(λ). The total isotropic photon emission, I_E, as a function of Q_Y, the fluorescence quantum, is given then by
• 
  I_E=Q_YA≈Q_YF_Xa₀(λ)N  (2)
• The major function of a fluorescent-based biosensor is to measure N using I_Ebased on equation (2) in the presence of F_X. Although F_Xgenerally has a slightly different wavelength from I_E, it is typically 4-5 orders larger and therefore needs to be blocked during detection.
• In one embodiment, a low-power diode pumped solid-state (DPSS) green laser is used with an output wavelength of 532 nm to create F_X. To block F_Xfrom reaching integrated photo-detector104, a multi-layer thin film dielectric long-pass (edge) optical filter (discussed below in connection withFIG. 3) is designed and fabricated. In one embodiment, the optical filter comprises 20 layers of ZnS (n=2.30) and Na₃AlF₆(n=1.35) with an overall thickness of 2.1 μm. The measured transmittance of the filter is illustrated inFIG. 2. The filter rejects F_Xby 98 dB at 532 nm, while having approximately 1 dB loss in the optical passband.
• In connection with the filter discussed above,FIG. 2 is agraph200 of the filter transmittance (in dB)201 and photodiode quantum efficiency (QE)202 versus the wavelength (in nanometers)203 in accordance with an embodiment of the present invention. Referring toFIG. 2,FIG. 2 illustrates the filter response which rejected (about −100 dB) the light that was below 532 nm (at the fluorescent excitation band) while passing light that was greater than 560 nm. The range of frequencies that are attenuated are referred to as “stopband” as shown inFIG. 2; whereas, the range of frequencies that can pass through the filter without being attenuated are referred to as “passband” as shown inFIG. 2. Further,FIG. 2 illustrates that the emitted light of 570 nm wavelength will not be attenuated very much thereby allowing the transducer and detection circuitry (discussed below in connection withFIG. 10) using the CMOS process to detect this light signal.FIG. 2 further illustrates the efficiency of the CMOS transducer. The plot on the right ofFIG. 2 represents how a transducer (e.g.,photodiode302 ofFIG. 3) efficiently converts light energy (e.g., photon energy) into electrical energy.
• In order to integrate the biochemical part of the assay with fluorescent detector104 (FIG. 1B), capturing probes106 (FIGS. 1C and 1D) are optically coupled to the CMOS detector using a fiber-optical faceplate (FOF) placed on top of CMOS integrated circuit101 (FIG. 1A) as shown inFIG. 3 in accordance with an embodiment of the present invention. Referring toFIG. 3, in conjunction withFIGS. 1A-D, in one embodiment, CMOSsensor array chip101 has a thickness of approximately 530 micrometers which includes asilicon substrate301 containing aphotodiode302 andreadout circuitry303 forming apixel304. On top ofsilicon substrate301 may include adielectric layer305, such as silicon dioxide, containing metal “curtains”306A-B. Curtains306A-B may collectively or individually be referred to as curtains306 or curtain306, respectively. Curtains306 may be employed for further optical shielding against optical crosstalk by taking advantages of integrated circuit structures in addition to physical distance between neighboring pixels. In one embodiment, curtains306 are composed of vias and metal layers (shown inFIG. 8A) and may encompass theentire photodiode302.Chip101 may include any number of curtains306 and the number of curtains306 shown inFIG. 3 is illustrative.
• FIG. 3 further illustrates an optical filter307 (long-pass filter discussed above) residing on the top ofchip101. In one embodiment,optical filter307 has a thickness of approximately 2.1 micrometers. On top oflong pass filter307 may reside a fiber-optical faceplate308. In one embodiment,optical filter307 includes layers of materials with a dissimilar refractive index.
• In one embodiment,optical filter307 is fabricated on the bottom of fiber-optical faceplate308 and on the top ofchip101.Optical filter307 prohibits light scattering and guides the two-dimensional fluorescence signals along the vertical direction of its fibers. In one embodiment, the thickness of fiber-optical faceplate308 may be between 0.5 millimeters and 3 millimeters which thermally isolates the 40-60° C. microarray assay fromCMOS chip101 and also creates adequate distance between the solution andchip101 without any significant signal loss. The exposed surface of fiber-optical faceplate308 may be polished glass (SiO₂)309 which is ideal for DNA capturing probe attachments using standard aldehyde-modified surfaces.
• As stated above, on top ofchip101 resides a thin film dielectricoptical filter307 which blocks the excitation light, while only taking emission light from fluorophores107 (e.g., Cy3 in this case) (indicated by arrows in silicon dioxide region305). As also stated above, on top offilter307 resides fiber-optic faceplate308, which brings the bottom surface image (in this biosensor case, transducers integrated within CMOS chip101) to the top surface where biological analytes will be spotted in this case. Fiber-optic faceplate308 may provide a good surface platform for DNAcapturing probe attachments106 while minimizing loss of signal due to the distance between the detectors (e.g.,chip101 on the very bottom) and the light generated byfluorophores107 in the biological analytes. The biological analytes can be spotted on the top surface of fiber-optic faceplate308 for the detection. As illustrated inFIG. 3, bio detection can be readily done on the single platform and the results are in digital numbers for further signal processing.
• FIG. 3 further illustrates capturinglayer106 used for capturing fluorescent labels that were excited by afluorescent excitation signal310 with a wavelength of approximately 532 nanometers.
• CMOS image sensors may use a process referred to as “direct integration” to measure the emitted light fromfluorophores107. In direct integration, the photocurrent generated inphotodiode302 is directly integrated (accumulated) on the photodiode capacitor. It is widely known in the art that direct integration can be carried out in an array format, where individual array components (i.e., pixels) measure light independently.
• In an alternative embodiment, a capacitive transimpedance amplifier (CTIA) is used to create the photocurrent integrator as illustrated inFIG. 4.FIG. 4 illustrates an embodiment of the present invention ofpixel304. Some of the components ofpixel304 are shown in block diagram form in order not to obscure the present invention in unnecessary detail. A person of ordinary skill in the art would understand the workings of these components.
• Referring toFIG. 4, in conjunction withFIG. 3,pixel304 includes a capacitive transimpedance amplifier (CTIA)401.CTIA401 includes aphotodiode302, anamplifier403 receiving an input voltage V_R1and acapacitor404 in the feedback mechanism ofamplifier403. Further,CTIA401 includes aswitch405.
• Unlike a direct integrator, the linearity ofCTIA401 is not limited by the photodiode junction capacitance voltage-dependency, and in addition, has a diode-independent well capacity set byfeedback capacitor404. In one embodiment,feedback capacitor404 is a 780 femtofarad poly-to-poly capacitor. Motivated by the advantages of CMOS digital-pixel-sensor (DPS) image sensors, an analog-to-digital converter406 (ADC) (e.g., 14 bit analog-to-digital converter) has been integrated within the pixels.ADC406 includes acomparator407 coupled to acounter408 where the output ofcounter408 is a 14-bit digital output.Comparator407 and counter408 are controlled by acontrol logic409 whose actions are coordinated by clock CLK2.Control logic409 may receivecontrol signals410 used to program the actions ofcontrol logic409.
• ADC406 compares the output ofCTIA401 with the external reference voltage, V_R2, to measure the time that the ramp reaches V_R2using comparator407. To suppress the offset ofcomparator407, a chopper stabilizedpreamplifier411 may be implemented with an overall voltage gain of 60 dB gain. Chopper stabilizedpreamplifier411 includesmixers412,413 as well asamplifiers414,415.Mixers412,413 may also be referred to as “choppers” or “modulators.”Mixer412 is the first chopper which modulates a signal to a higher frequency; whereas,mixer413 is the second chopper which demodulates the signal back to baseband with an offset to cancel the low noise.
• FIG. 4 illustratesphotodiode302 converting emitted light into electrical signal (e.g., current). Then,CTIA401 converts this current into voltage. Thefeedback switch405 inCTIA401 shown inFIG. 4 provides extra large resistance. In a very low current input case, there will be limited finite CMOS switch resistance that prohibits the output ofCTIA401 going from the desired voltage.Switch405 provides effectively much greater resistance than a single switch can provide while reducing the charge injections fromswitch405.
• Chopper stabilizedpre-amplifier411 is used to amplify the output ofCTIA401 while reducing 1/f noise of the 1^ststage of the amplifier used inpre-amplifier403 ofCTIA401. Anon-overlapping CLK generator416 is used to generate a clock signal to coordinate the multiplication of signal voltages bymixers412,413. The actions ofCLK generator416 are coordinated by clock CLK1.
• The time that it takes for the voltage output ofCTIA401 to reach a certain voltage provides information about the amount of light detected byphotodiode302. In turn, the amount of light detected byphotodiode302 is proportional to the abundance of the target analytes on the surface of fiber-optical faceplate308 (FIG. 3). In one embodiment, all entire systems aforementioned are integrated in a single pixel.
• FIG. 5 is agraph500 illustrating the measured sensitivity of the integrated microarray and the background current (I_dc) present in the system as a function of temperature in accordance with an embodiment of the present invention. Referring toFIG. 5,graph500 illustrates the measured sensitivity of the integrated microarray (as current in Amps)501 as a function of the incident photon flux502 (λ=532 nm), incident signal power503 (λ=532 nm), and the estimated surface density of the fluorescent labels504 (Cy3 label with a 3 mW excitation source). The measured dark current of each pixel304 (FIG. 3) is approximately 12 fA at room temperature, which is much lower than the measured signals and therefore sufficiently low for typical fluorescence spectroscopy applications.FIG. 5 illustrates the sensitivity of a transducer (e.g.,photodiode302 ofFIGS. 3 and 4) with CTIA401 (FIG. 4).FIG. 5 further illustrates the “dark current” (indicated as “I_dc” inFIG. 5) that was measured at different temperatures (e.g., 20° C., 30° C., 40° C.). The dark current is the unwanted current that is typically generated by thermal agitation and leakage in photosensitive devices, such as photodiodes and charge-coupled devices. The integration time (indicated as “T_int”) indicates the time that was set inpixel304 ofFIG. 4 to measure the currents ofFIG. 5.
• In addition to conventional microarray applications, system100 (FIGS. 1A-D) is capable of real-time detection of fluorescent signals emitted from biological samples in solution.FIG. 6 is agraph600 illustrating the measured real-time DNA hybridization kinetics for three exemplaryDNA target concentrations601,602,603 in solution in accordance with an embodiment of the present invention.
• Referring toFIG. 6, thetarget strand5′-AGCAACATTTTGCTGCCG-3′ is labeled with Cy3 and theprobe strand3′-TCGTTGTAAAACGACGGC-5′ is labeled with a black-hole quencher. The sequences are complementary so that binding occurs.
• DNA target concentration601 corresponds to a 0.5 nmol target with a k_H(normalized inverse of time constant) equal to 0.43×10⁻³sec⁻¹. k_Hrefers to the normalized inverse of time constant which shows how much bindings occur in a given time. For example, the 0.5 nmol target case has more bindings in comparison to the 0.125 nmol target case in a given time.DNA target concentration602 corresponds to a 0.25 nmol target with a k_Hequal to 0.31×10⁻³sec⁻¹.DNA target concentration603 corresponds to a 0.125 nmol target with a k_Hequal to 0.22×10⁻³sec⁻¹.
• The results shown inFIG. 6 confirm that the measured rate of capturing is directly proportional to the analyte concentration. In one embodiment, the biosensor can detect hybridization kinetics in real-time. In a conventional microarray, washed and dried analytes labeled with fluorophores are typically used. Therefore, the researcher should wait until hybridization of DNA is done. Thus, only a steady-state response can be detected. However, in system100 (FIGS. 1A-D), the solution can be spotted on the top of the sensor so that the detection can be performed during hybridization. Therefore, this integrated biosensor is able to detect real-time hybridization kinetic behavior. It is noted that this real-time detection does not require biosensors to be washed and waited until biological analytes get saturated. Detecting hybridization kinetic also provides information about the abundance of target analytes. In the illustrative case, all the capturing probes are labeled by fluorophores (e.g., Cy3) and the targets are labeled by quenchers, which extinct the light generated by fluorophores. The normalized results are shown inFIG. 6. The largest amount of target analytes is 0.5 nmol with the smallest time constant. FIG.6 illustrates that larger target analytes in the solution represent the faster rising response. In summary, this type of integrated biosensor can be used for not only conventional detection (steady-state), but also real-time hybridization kinetics (transient).
• The integrated biosensor of system100 (FIGS. 1A-D) includes a fully integrated fluorescent-based microarray system. The achieved performances of this system in terms of sensitivity, compactness, versatility, and cost, satisfy the requirements of many biotechnology applications beyond DNA microarrays.
• Though fluorescent based detection methods have been popularly used with microarrays, label-free detection of analytes using their intrinsic properties (e.g., charge, mass, absorption spectra) has generated a lot of interest in the research community. Label-free detection offers several advantages such as reduction in cost, omission of the molecular labeling process, feasibility of real time detection and ease of integration with standard CMOS processes.
• Among the various techniques for label-free detection, impedance spectroscopy-based detection is the most compatible with current silicon-based very-large-scale integrated (VLSI) systems and integrated electronics. The concept behind this method is illustrated inFIG. 7 in accordance with an embodiment of the present invention. In this method, anelectrode surface701 is first functionalized by immobilizing probes702 (indicated as Y's attached to top of electrode surface701) on theelectrode surface701.Electrodes701 are placed in anaqueous solution703 containing analytes704 (analytes704 are depicted inFIG. 7 as various symbols, such as circles and triangles). Any number ofanalytes704 may be present inaqueous solution703. Whenanalytes704 bind toprobes702, the physiochemical characteristics of the electrode-electrolyte interface changes which subsequently results in changes in the impedance of the interface. It is known in the art that such changes in the impedance of the interface can be measured using an electronic sensor and the results can be related to the binding events, their frequency, and their abundance. As illustrated inFIG. 7, there are three important terms contributing to the impedance: R_b, the resistance of the solution and the impedance of the double layer represented by R_dand C_d. On the right ofFIG. 7, the R_b′, R_d′ and C_d′ indicate the modified values due to analyte binding. The spectroscopy-based detection is known as “impedance” spectroscopy-based detection to signify the interest in the changes in both the resistance and the capacitance of the interface as a function of excitation frequency. As impedance is a purely electrical quantity, one of the key advantages of this method is that all the signals are completely in the electronic domain. This attractive feature enables straightforward integration with standard CMOS processes.
• In an illustrative design using the principles of the present invention, a 10×10 array of impedance sensors is integrated with the detection circuitry (shown inFIG. 10) for sensing the impedance of the interface betweenelectrode701 and electrolyte. Some innovative aspects of the design include the use of on-chip sensing electrodes and simple detection circuitry which is present underneath each ofelectrodes701, permitting large-scale integration and high packing density. Another innovative aspect is that the whole chip, including the sensors, may be fabricated using 0.35 μm standard CMOS process. Since CMOS process is widely used in digital integrated circuits, large scale production of “integrated impedance sensors” could lead to significant lowering of costs.
• FIGS. 8A-B illustrate apixel800 with on-chip electrode701 (FIG. 7) that is formed by creating anopening801 in apassivation layer802 covering the metal layers plusdielectric layer803 in accordance with an embodiment of the present invention. Referring toFIGS. 8A-B, in conjunction withFIG. 7, in a CMOS process,active devices804 are fabricated on asilicon substrate805, and 3-4 metal layers, which are used to route various signals, are stacked on top of theseactive devices804.FIG. 8A further illustratesvias806 used to interconnect various layers.Passivation layer802 covers the top layer ofmetal layers803 in order to protectmetal layers803 anddevices804 underneath from the outside environment. Anopening801 inpassivation layer802 exposes the top layer ofmetal layers803 to an aqueous solution of affinity-based biosensors, which is used as thesensing electrode701. Capturingprobes702 are attached tosensing electrode701. In one embodiment,passivation layer802 has a thickness of approximately 2 micrometers.
• FIG. 9 illustrates a scanning electron microscope (SEM)photograph900 of a surface of the fabricated pixel800 (FIGS. 8A-B) in accordance with an embodiment of the present invention. Referring toFIG. 9, in conjunction withFIGS. 7-8, in one embodiment,electrode701 is roughly of the size 40 μm×40 μm. In one embodiment, the edges ofelectrode701 are rounded off in order to approximate a circle since a drop placed on top ofelectrode701 tends to take a circular shape. In one embodiment,electrodes701 are spaced approximately 50 μm apart. In one embodiment, the die area is approximately 2 mm×2 mm.
• FIG. 10 illustrates the excitation scheme of achip1000 in accordance with an embodiment of the present invention. Referring toFIG. 10, in conjunction withFIGS. 7 and 8, the surface ofelectrodes701 is functionalized with capture probes702.Solution703 containinganalytes704 is dropped on top of adetection circuitry1001 residing onsilicon substrate805 ofchip1000. In measuring the impedance of the system, an excitation source1002, which is a sinusoidal function generator, is used to generate sine waves in the frequency range from a 0.01 Hz to 500 MHz. In one embodiment, excitation source1002 typically generates sine waves in the frequency range between 10 Hz to 10 MHz. The excitation signal may be applied to an Ag/AgCl electrode1003 which is dipped insolution703. The impedance of the interface may be measured bydetection circuitry1001.
• FIG. 11 illustrates the basic concept used in measuring impedance by detection circuitry1001 (FIG. 10) in accordance with an embodiment of the present invention. Referring toFIG. 11, the input sinusoid Vin=A cos ωt1101 is applied as the excitation source. Theimpedance Z1102 represents the sum of the solution resistance and the interface impedance. The current flowing through Z is given by
• $\begin{array}{cc}I=\frac{V}{Z}=A1\cos \left(\omega t-\theta \right)& \left(3\right)\end{array}$
• This current is multiplied by two sinusoidal signals B cos ωt1103 and B sin ωt1104, which have the same frequency of the excitation source. The path in which the multiplication with B cos ωt is calledI path1105. The other path is generally referred to as theQ path1106. After multiplication bymultipliers1107,1108 and subsequent low pass filtering bylow pass filters1109,1110, a signal proportional to cos θ is generated inI path1105, while the signal proportional to sin θ is generated inQ path1106. Using these two signals, it is possible to calculate both the magnitude and phase of the complex impedance denoted by Z inFIG. 11.
• FIG. 12 illustrates acircuit1200 for the each individual pixel800 (FIGS. 8A-B) in accordance with an embodiment of the present invention. Referring toFIG. 12, in conjunction withFIGS. 7,8 and11,circuit1200 includes a commongate input stage1201 which includes a p-type transistor1202 coupled to n-type transistors1203,1204. The gate of n-type transistor1203 is coupled to adifferential amplifier1205.Electrode701 is connected to inputnode1206 ofcommon gate amplifier1205.Electrode701 is coupled to the input voltage (V_in). The gate of n-type transistor1204 is biased with a signal labeled “vbias_tail.” The negative input ofcommon gate amplifier1205 is coupled tonode1206; whereas, the positive input ofcommon gate amplifier1205 receives an input voltage labeled “vin_cm.” The source of p-type transistor1202 is coupled topower supply1207; whereas, the source of n-type transistor1204 and V_inare coupled toground1208.
• Circuit1200 additionally includes amixer1209. Front-end amplifier1201 andmixer1209 may be implemented on-chip while other components (e.g., low pass filter and signal processing blocks) may be implemented off-chip in order to reduce the area ofpixel800.
• In one embodiment, commongate input stage1201 presents a low input impedance below 100 ohms for the entire frequency range from DC to 50 MHz. This is achieved by further reducing the transconductance ofinput transistor1202 using a simpledifferential amplifier1205 as the gain-boosting circuitry. Another important function ofdifferential amplifier1205 is that it helps to set the DC potential atelectrodes701. This helps in maintaining zero DC potential between the working electrode common to all the electrodes and the on-chip electrode. The current flowing throughinput stage1201 is transferred usingcurrent mirrors1210 to a set of double-balanced Gilbert-cell mixers1211. Double-balancedGilbert cell mixers1211 are adopted to suppress the component at the signal frequency caused by I and Qsquare waves1105,1106 applied tomixer1209. To minimize mismatch, an exact replica of input circuit1201 (circuit1212) is used carrying the same amount of current but with no input being applied. The only exception is that differential amplifier1213 used inreplica bias circuit1212 has lesser current. Furthermore,resistors1214,1215 are used in the load ofmixer1209 in order to minimize the flicker noise (1/f noise) atoutput1216.
• Mixer1209 additionally includes a current source identified by “Inix” that is inputted tocurrent mirrors1210. The gate of thetransistors1217,1218 ofcurrent mirrors1210 receives an opposite input voltage, vin+ and vin−, respectively. The drains oftransistors1217,1218 are coupled to the sources oftransistors1219,1220,1221,1222 ofmixers1211. The gates oftransistors1219,1222 receive a positive oscillator voltage (identified as “VLO+”); whereas, the gates oftransistors1220,1221 receive a negative oscillator voltage (identified as “VLO−”). The negative and positive output voltage of mixers, identified as “Vout−” and “Vout+,” respectively, are coupled toresistors1214,1215, respectively. Resistors1214,1215 are coupled toground1208.
• As discussed above,circuit1212 is a replica ofinput circuit1201.Circuit1212 includes a p-type transistor1223 coupled topower supply1207. The gate oftransistor1223 is coupled to the gate oftransistor1217. The drain oftransistor1223 is coupled to the source of n-type transistor1224. The gate oftransistor1224 is coupled to differential amplifier1213 with its negative input coupled to the source oftransistor1224. The positive input of differential amplifier1213 receives the input voltage labeled “vin_cm.” The source oftransistor1224 is coupled to a current source labeled “lcg” which is coupled toground1208.
• The pixel level performance metrics using the components ofcircuit1200 are shown in Table I.

• TABLE I
  Performance metrics ofpixel 304

  Gain of the cell	90dB
  Bandwidth
  	10 Hz-50 MHz
  Noise at the output referred to the	<0.4 nA rms over 100 Hz bandwidth
  input current
  Input impedance	<100 ohm from dc to 50 MHz at all
  	corners
  Max input current	20 μA
  Current consumption	210 μA at 3.3 V supply

• In one embodiment,circuit1200 has a current to voltage gain of 90 dB and the input referred noise current is less than 0.2 nA rms for 100 Hz bandwidth. In one embodiment, eachpixel800 consumes 210 μA of current with a 3.3V power supply. In one embodiment, the maximum input current for the circuit remains linear at 20 μA.
• Although the systems are described in connection with several embodiments, it is not intended to be limited to the specific forms set forth herein, but on the contrary, it is intended to cover such alternatives, modifications and equivalents, as can be reasonably included within the spirit and scope of the invention as defined by the appended claims.

Claims (24)

1. A biosensor system, comprising:

a complementary metal-oxide-semiconductor integrated circuit, wherein said complementary metal-oxide-semiconductor integrated circuit comprises a silicon substrate and a dielectric layer on said silicon substrate;

an optical filter fabricated on said complementary metal-oxide-semiconductor integrated circuit; and

a plurality of capturing probes optically coupled to said complementary metal-oxide-semiconductor integrated circuit.

2. The biosensor system as recited inclaim 1, wherein said optical filter prohibits light scattering.

3. The biosensor system as recited inclaim 1, wherein said optical filter guides two-dimensional signals along a vertical direction of its fibers.

4. The biosensor system as recited inclaim 1 further comprising:

a fiber-optical faceplate positioned on said optical filter, wherein said plurality of capturing probes are optically coupled to said complementary metal-oxide-semiconductor integrated circuit using said fiber-optical faceplate.

5. The biosensor system as recited inclaim 4, wherein a thickness of said fiber-optical faceplate is between 0.5 millimeters and 3.0 millimeters.

6. The biosensor system as recited inclaim 4, wherein an exposed surface of said fiber-optical faceplate is silicon dioxide.

7. The biosensor system as recited inclaim 1, wherein said complementary metal-oxide-semiconductor integrated circuit comprises a biosensor pixel fabricated in said silicon substrate, wherein said biosensor pixel comprises:

a capacitive transimpedance amplifier used to create a photocurrent integrator, wherein said capacitive transimpedance amplifier comprises a photodiode.

8. The biosensor system as recited inclaim 7, wherein a photocurrent generated in said photodiode is directly integrated on a photodiode capacitor.

9. The biosensor system as recited inclaim 7, wherein said biosensor pixel further comprises:

an analog-to-digital converter configured to compare an output of said capacitive transimpedance amplifier with an external reference voltage using a comparator.

10. The biosensor as recited inclaim 9, wherein said biosensor pixel further comprises:

a chopper stabilized preamplifier configured to suppress an offset of said comparator.

11. The biosensor as recited inclaim 1, wherein said optical filter comprises a plurality of layers of materials with dissimilar refractive index.

12. The biosensor as recited inclaim 1, wherein said optical filter is configured to block an excitation signal to said complementary metal-oxide-semiconductor integrated circuit.

13. The biosensor as recited inclaim 12, wherein said excitation signal has a wavelength of 532 nanometers.

14. The biosensor as recited inclaim 12, wherein said excitation signal is used to excite fluorescent labels.

15. A biosensor system, comprising:

a silicon substrate;

active devices fabricated on said silicon substrate;

a plurality of metal layers stacked on top of said active devices;

a passivation layer covering a top metal layer of said plurality of metal layers in order to protect said plurality of metal layers, wherein said passivation layer comprises an opening configured to expose said top metal layer, wherein said opening is used as a sensing electrode; and

a plurality of probes attached to said sensing electrode.

16. The biosensor system as recited inclaim 15, wherein said sensing electrode is placed in a solution containing analytes.

17. The biosensor system as recited inclaim 16 further comprising:

an interface between said sensing electrode and an electrolyte, wherein an impedance of said interface is changed when an analyte of interest binds to one of said plurality of probes.

18. The biosensor system as recited inclaim 17, wherein said change in said impedance of said interface is measured using an electronic sensor.

19. The biosensor system as recited inclaim 17, wherein said change in said impedance of said interface is measured using an integrated circuit.

20. The biosensor system as recited inclaim 19, wherein said impedance is measured by de-coupling an excitation signal into a first and a second path.

21. The biosensor system as recited inclaim 20, wherein current traveling through said interface is multiplied prior to traveling through said first and said second path.

22. The biosensor system as recited inclaim 17 further comprising:

impedance sensors for sensing said impedance of said interface, wherein said impedance sensors are integrated in an array format.

23. The biosensor system as recited inclaim 15 further comprising:

an excitation electrode, wherein said excitation electrode is an Ag/AgCL electrode.

24. The biosensor system as recited inclaim 23, wherein an excitation signal is applied to said Ag/AgCL electrode, wherein said excitation signal is generated using a sinusoidal function generator generating sine waves in a frequency range from 0.01 Hertz to 500 Megahertz.

Images