Dynamic BIONEMS sensors and arrays of BIONEMS sensor immersed in fluids

Dynamic BIONEMS sensor and BIONEMS sensor array immersed in liquid

Related Application This application is related to the US provisional patent application serial number 60/379,710 filed on May 7, 2002; serial number 60/379,660 filed on May 7, 2002; serial number 60/379,645 filed on May 7, 2002 ; Serial number 60/379,552 submitted on May 7, 2002; serial number 60/379,711 submitted on May 7, 2002; serial number 60/379,543 submitted on May 7, 2002; submitted on May 7, 2002 The serial number 60/379,643; the serial number 60/379,708 submitted on May 7, 2002; and the serial number 60/379,681 submitted on May 7, 2002 are related. They are hereby cited for reference and are required by 35 USC 119 Its priority.

The included co-pending applications know that this application will simultaneously submit the serial number (PAU.34) titled "An Apparatus And Method ForVacuum-Based Nanomechanical Energy, Force, And Mass Sensors" and the serial number titled "A Method And Apparatus" For Providing Signal Analysis Of A Bionems Resonator" application serial number (PAU.35) is incorporated by reference in its entirety as given here. In addition, this application combines the U.S. Patent Application Serial No. 10/138,538, filed on May 3, 2002, entitled "An Apparatus and Method for Ultrasensitive Nanoelectrochemical Mass Detection"; and the U.S. Patent Application Serial No. 10/138,538 filed on August 9, 2001, entitled "Active NEMS Arrays". ForBiochemical Analyses" US Patent Application Serial No. 09/927,779 is incorporated by reference as given in its entirety.

Technical field

The invention relates to the field of liquid bioNEMS devices and methods of operation.

Background technique

In recent years, many advances have been made in the fields of NEMS and Chemical Force Microscopy (CFM). The NEMS method produces a series of cantilever with small length and thickness, high Q and high frequency resonance (cantilever). When working under ideal conditions (low T, vacuum), these NEMS devices exhibit unprecedented sensitivity. At the larger size levels (AFM, CFM), a variety of tasks involve analyzing the forces exerted by the interactions between individual molecules, from hydrogen bonds and antibody-antigen interactions to covalent bonds. The AFM cantilever decorated with biomolecules and interacting with the derivatized surface or derivatized magnetic beads proves that the antigen-antibody interaction exhibits a force of the order of 100 pN, and the covalent bond exhibits a force of about 1-10 nN. These watershed experiments have shown that chemical activity at random limits can be measured, but it has also shown that it is difficult to realize this potential in a small, portable, and rugged device.

According to the present invention, a method is needed to reduce the size of the cantilever relative to the size of the NEMS, provide the required transient response, small size, and sensitivity to a single molecule, which is required to build the device into a single unit performance. Of course, placing the NEMS cantilever in the solution and keeping it at room temperature will need to modify the detection strategy commonly used by CFM or NEMS. The liquid will damp the NEMS arm, making it impossible to perform resonance detection, and the heat energy of the solution will impact the cantilever.

For these potential difficulties, a method needs to be developed as the main part of the experiment.

Unlike the traditional CFM, a method for measuring the force of a single (or a small amount) chemical bond without recording the deflection of the cantilever is required according to the present invention.

There is a need for some type of NEMS cantilever design that detects the presence of chemical bonds by using an integral sensor to limit the cantilevers large thermally driven motion.

According to the present invention, there is also a need for a device using a BioNEMS cantilever array with different chemical decorations on the system, while providing high reliability and sensitivity to concentration.

In addition, according to the present invention, some method is needed to translate fluctuating "noise" into signals, and possible biology for assembling and employing useful and robust experiments.

Microarray technology has significant advantages in analyzing protein receptors and their ligands, and analyzing gene expression distribution. For example, microarrays of thousands of objects have become a major technology used in the pharmaceutical development industry. These micro-arrays are produced by photolithography, micro-punching or micro-dotting, and dot arrays (20-100 planes) are produced on the substrate. Usually, the array is read out by covering the array with fluorescently labeled analytes, and scanning the array with a microfluorometer to determine the amount of binding. Although these methods are becoming more and more common, the large-size readout instruments and the inherent limitations of the fluorescence analysis used make them completely unsuitable for applications that require portability and robust performance at the same time. In addition, they are single-use devices, so they are not easily suitable for applications that require continuous monitoring. Finally, these devices rely on a large number of analytes, making them unsuitable for the most powerful recent advances in drug discovery through combinatorial chemistry, or for the most sensitive gene expression tests.

Another purpose of the research is to develop a new nano-scale biochip technology (BioNEMS) that can detect the binding of a single biomolecule to its receptor. Numerous chemical force microscopy (CFM) documents indicate that the modified AFM is suitable for measuring the binding force of interactions ranging from single hydrogen bonds and single acceptor ligand interactions to single covalent bond interactions. The range of these forces is just within the detection capability of the AFM instrument; however, the AFM cantilever in solution does not have the transient response required to allow the binding and debinding of the biological ligand to its receptor to proceed reliably. Perhaps the most important is the large size of the device required to perform AFM/CFM, and the well-known sensitivity of AFM to airborne and surface vibrations.

There is a need for a technique that is as successful as CFM in detecting single-molecule interactions, but down to the NEMS level, allowing it to respond quickly enough to subsequent binding and unbinding activities. Given the size of the chemical force, the most robust working mode of BioNEMS is not to directly measure the binding force. According to the present invention, a device is required that uses the current fluctuations in the position of the NEMS cantilever and then uses the integral sensor without the support device used in AFM.

Summary of the invention

According to the present invention, the movement of the NEMS cantilever will be used to follow the bound and unbound activities. The basic idea is that a cantilever without an acceptor ligand pair at the end will fluctuate more significantly in its position than a cantilever constrained by a ligand-acceptor pair. The strong ligand-acceptor bond can partially prevent the cantilever motion for a long period of time (~ton for the ligand-acceptor pair); weaker interaction will change the statistical results of the cantilever motion.

The NEMS device reduced to a small size yields a number of significant advantages. "NEMS" is used in this description to mean at least one device with a size equal to or less than one micron. It does not exclude the possibility that the "NEMS" device may have one or more other sizes larger than one micron. In addition, as can be appreciated, there is generally no significant difference between the performance of a device with a size of one micron or less and a device larger than one micron in size. The more significant importance of the term "NEMS" is that it is suspected that the device shares certain properties with devices down to sub-micron size, or shares properties that are unique to sub-micron devices or operations. As already proposed, the small size of the NEMS device allows it to make a more pronounced response to bound and unbound kinetics. This high-frequency response is very critical to follow the randomness of the receptor ligand interaction. Most receptor-ligand pairs interact dynamically, bind, and maintain the binding time from microseconds to seconds (depending on the actual The receptor-ligand pair), and then released. If the experiment is to track biomolecular interactions, the high frequency response (~MHz) is very important. The ability to solve the opening and closing of various membrane channels in patch-clamp (giga-ohm sealing) technology has completely changed our understanding of the physical and biochemical foundations of nerve function; before patch damping, experiment Just try to decode the molecular mechanism by recording a large number of thin film channels. We believe that the analysis of biomolecules is currently in the same state, and is limited by the large amount of materials required and the time tailing inherent in even the most sensitive experiments. Therefore, the present invention carefully considers BioNEMS, and truly puts our biomolecule analysis to the limit of randomness.

One effect of this method is to use the thermal motion of the cantilever as the driving force (usually the main limitation in AFM). In addition, when the size of the cantilever is reduced, the noise of the movement of the cantilever becomes smaller. In addition, the small size of the NEMS device allows the construction of a detector array (500 cantilever) in a small effective volume (100pL). The latter advantage is extremely important because it makes it possible to sense the second messengers present in RNA levels, proteins and individual units.

The BioNEMS method of the present invention greatly reduces the size and properties required for instrument operation (compared to AFM/CFM). The sensor for the cantilever movement will be integrated with the NEMS cantilever, which eliminates the size and density limitations of the cantilever movement optical detection used in AFM. This will allow the BioNEMS cantilever to be smaller and more compactly assembled compared to the actual situation in AFM.

As described in the section describing NEMS sensing below, the integrated piezoresistive sensor will provide much greater sensitivity than is required to record the movement of the NEMS cantilever in liquid water. As a result, with proper integration, the sensors, the detectors needed to follow the movement of the cantilever, the logic needed to translate the movement, and the circuits needed to communicate the results can be packaged into one device. Unlike the hint given by its name, the current "DNA chip" or "Proteomics chip" technology requires a large and cumbersome reader to translate the combination of chemicals into a sensor package that is a few centimeters in length and width. The BioNEMS method outlined here promises that the package size is consistent with the term "chip" (~DIP size), resulting in a variety of applications that are impossible or impractical for other methods.

The purpose of the proposed work is to use the thermally driven movement of the cantilever and its modulation through the receptor-ligand interaction. The movement of the cantilever will be explained using knowledge of the physics and random biochemistry of the NEMS cantilever in solution. Therefore, the structure of this type of BioNEMS in use requires researchers dedicated to the manufacture of NEMS, biologists dedicated to the biochemical improvement of the device surface, physicists interested in the fluid dynamics of the NEMS device, and complete extraction and extraction from experiments. The analytical information scientists work closely together.

The BioNEMS research efforts described here have multiple driving effects ranging from basic principles of applied science to the development of new nanoscale fluid technology. Our purpose is to study, understand, and improve the technology of constructing BioNEMS, and then explain its novel uses.

Examples include: Basic research on the properties of NEMS in solution Basic research on single molecule chemistry Cytological research on hormones, growth factors and second messengers. For direct analysis using traditional techniques, the growth factors released from the cells are usually very low in concentration and too small.

Use BioNEMS as a sensor for the output of combinatorial chemical synthesis in the drug discovery work.

Used as a sensitive "gene chip" for detecting DNA sequences, or as a biohazard sensor.

Used as a monitor of environmental mycin concentration.

The present invention is limited to a sub-micron bioNEMS device, including a stent and a piezoresistive cantilever coupled with the stent and extending therefrom, having a length of 1, a width of w, and an end, wherein the cantilever has a width reduced to b and a length of The restricted part of l1, and the biologically functionalized part at or near the end. The restricting part is composed of a plurality of pins whose width is reduced to b and connected to the bracket. It is best to set up two pins and separate them by a distance of w-2b.

The bioNEMS device also includes a bias current source applied to the cantilever restriction part, and the magnitude of the bias current is limited by the maximum acceptable temperature rise at the end of the biofunctionalization. The maximum acceptable temperature rise at the end of the biofunctionalization is approximately 1 degreeK.

The present invention is also an improvement of the piezoresistive bioNEMS device immersed in liquid, which includes at least one vibrating cantilever with a length of 1, and the amplitude is selected to minimize the background Johnson noise relative to the signal intensity generated by the piezoresistive bioNEMS device. In one embodiment, the signal strength is based on the thermomechanical noise of the piezoresistive bioNEMS device in the liquid. The piezoresistive cantilever has a width w, and a restricted portion whose width is reduced to b, where the reduced width b is selected to reduce Johnson noise relative to the signal strength generated by the piezoresistive bioNEMS device.

The present invention is also characterized by an improvement of a biologically functional bioNEMS device immersed in a liquid, which includes a receptor arranged on the bioNESM device for binding to a ligand of interest, and a bioNESM device arranged on the bioNESM device with the receptor. A catalyst that enhances the binding rate coefficient of the receptor and the ligand of interest. The catalyst reduces the receptor-ligand binding activity. In one embodiment, the receptor is designed by forced evolution to preferentially bind to the ligand of interest.

The present invention also relates to a sub-micron device, including a carrier signal source, a support, a piezoresistive cantilever coupled to the support and extending therefrom, and an element arranged on the cantilever and electromagnetically coupled to the source, so as to pass from all sources. The carrier signal from the source drives the cantilever. The element includes a magnetic film provided on a cantilever, and the source generates electromagnetic waves coupled with the magnetic film.

The present invention also relates to a device including a plurality of NEMS resonators or sensors, each NEMS sensor generates an output signal; and a device or circuit for processing a plurality of corresponding output signals generated by a plurality of NEMS sensors to obtain a collective output signal . The device averages multiple output signals, so that the aggregate output signal is an average value. The device judges whether a predetermined part of a plurality of output signals exceeds a threshold value within a predetermined time window. Each of the plurality of NEMS sensors is biologically functional, and the device can effectively increase the ligand capture rate by generating only a collective output signal indicating an increase in ligand capture rate compared with a NEMS sensor.

The present invention also relates to a device working in liquid, including a plurality of NEMS sensors immersed in the liquid to form adjacent sensor arrays, each NEMS sensor generates an output signal, and the movement of two adjacent NEMS sensors passes through the adjacent NEMS The immersed liquids are coupled to each other. The motions of the first and second NEMS sensors are correlated with each other, and include two adjacent NEMS sensors, C12=<x1(0)x2(t)>, as defined by the following formula: ddtC12(t)=-kBTX12 (t)----t>0]]> where kB is Boltzmann's constant, T is the temperature of the liquid, t is the time, and X12 is the "magnetization coefficient", which gives the Force F1, the displacement of the second sensor x2(t), which defines the overall mean value of the position of the second sensor by the following formula: <x2(t)>=-X12(tt)F1( t)dt]]>The present invention is a device that works in a liquid containing a microfluidic flow channel, wherein the microfluidic flow channel carries liquid and includes at least one NEMS arranged in the microfluidic flow channel Sensor, so as to sense the properties of the liquid through the NEMS sensor. The NEMS sensor is biologically functionalized, and the properties sensed by the NEMS sensor are present or absent in the ligand liquid that the NEMS sensor has been biologically functionalized.

The device also includes a plurality of NEMS sensors, and each sensor is jointly arranged in the liquid flow channel. The device also includes multiple fluid flow channels in which multiple NEMS sensors are redistributed. Multiple NEMS sensors are either surface-fabricated or thin-film fabricated.

The present invention includes a method for manufacturing a bioNEMS device from a thin film, including the step of providing a heterostructure, wherein the heterostructure includes a wafer layer, an etch stop layer on the wafer layer, a NEMS device layer on the etch stop layer, and Piezoresistive layer on the NEMS device layer. A trench is etched through the wafer layer to the etch stop layer, defining the area that will become the thin film that defines the NEMS device. Remove the etch stop layer from the bottom of the trench to the device layer. Conductive contact points are selectively formed on the piezoresistive layer of the film by electron beam lithography. The area to be biofunctionalized is selectively formed on the piezoresistive layer of the film by electron beam lithography. NEMS devices are selectively formed on the piezoresistive layer of the thin film containing the area to be biofunctionalized by electron beam lithography. The thin film is selectively plasma etched to remove the unmasked parts to define the suspended NEMS device. Liquid flow channels are selectively cast in the layer of elastic material deposited around the thin film. Biofunctionalize the selected area on the NEMS device.

The step of forming the heterostructure further includes polishing the wafer layer to improve the adhesion of the elastic material layer and to make the wafer layer thinner. The step of selectively plasma-etching the film to remove the unmasked portion and constraining the suspended NEMS device includes selectively etching off the unmasked portion of the NEMS device layer by the vertical plasma. The step of selectively pouring a liquid flow channel in the elastic material layer arranged around the film includes selectively setting a photoresist layer to define the liquid flow channel, and disposing an elastic material layer on the selectively arranged photoresist layer , And remove the photoresist layer to define liquid flow channels.

The present invention discloses a NEMS device working in liquid, including a resonance element whose end is immersed in the liquid, a binding molecule attached to the end; and a pompons attached to the binding molecule, which provides damping to the dissipation noise applied to the element from the liquid force.

The present invention also includes a method of operating the aforementioned NEMS device.

Although the device and method are described or will be described with functional interpretation for grammatical flexibility, it is understood that unless explicitly stated under 35 USC 112, the claims are not interpreted as restrictive interpretation by "device" or "step" in any respect. Make necessary restrictions. However, under the judicial meaning of the equivalent, it is consistent with the defined meaning provided by the claims and the full scope of its equivalents, and the claims described in the following table under 35 USC 112 are consistent with the full legal equivalents under 35 USC 112. The present invention will be more vividly understood by referring to the accompanying drawings, in which the same reference numerals denote the same elements.

Description of the drawings

Figure 1 is a graph showing the increase in intensity of liquid-coupled thermomechanical noise relative to Johnson noise when all other dimensions are fixed.

Figure 2 shows the curve of liquid damping thermo-mechanical noise.

Figure 3 is a graph of expected signals in several liquids for a bias current of 250μA, for b=0.6μm, t=130nm, w=2.5μm, l=15μm and l1=0.6μm cantilever.

Fig. 4 is a schematic side view of the cantilever microscopic magnification scale.

Figure 5 is a graph of partial receptor occupancy as a function of non-specific binding activity.

Figure 6 shows a detector called a "phase detector" or "lock-in amplifier", or related receiver.

Figure 7 shows a graph of the detector performance Pd as a function of SNR.

Figure 8a is a schematic top view of a dual cantilever system.

Figure 8b is a schematic side cross-sectional view of the system of Figure 8a.

Fig. 8c is a schematic top view of another embodiment, in which the cantilever is a coupling base.

Figure 9 is a graph of the velocity component parallel to the oscillating cantilever as a function of distance r.

Fig. 10 is a cross-sectional schematic side view of a piezoresistive cantilever coupled with a microfluidic flow channel.

Figures 11a-11c are perspective views in which the magnification of the cantilever array shown in Figure 10 has been increased.

Figure 12a is a scanning electron microscope photo of the bioNEMS sensor. Fig. 12b is a top view of the sensor of Fig. 12a.

Figures 13a-13m are a series of diagrams showing a method of manufacturing a bioNEMS fluid sensor from a thin film.

Fig. 14 is a diagram simulating the dynamic action of a cantilever with a molecular bond energy pom-pom as an additional damper.

The present invention and various embodiments can be better understood by referring to the detailed description of the following preferred embodiments, which are taken as illustrative examples of the present invention defined in the claims. It can be understood that the present invention defined by the claims can be broader than the illustrative embodiments described below.

detailed description

The illustrated embodiment relates to the restriction of the bias flow that can be applied to the piezoresistive BioNEMS device in the liquid. As long as the responsivity is proportional to the bias current, R=/G, the obtainable force sensitivity depends on the maximum allowable bias current value. The maximum actual bias current value is determined by the maximum temperature increase in BioNEMS deemed acceptable.

For the purpose of illustration, it is assumed that the bioNEMS sensor or the cantilever 10 shown in the perspective view of the photomicrograph Fig. 12a and the top view of Fig. 12b can be compared to having a "drive plate with a cutout on the bottom." However, it should be understood that the geometry of the sensor 10 is completely general and includes any type of cantilever, dual damping beam, blade or any other sub-micron vibration structure. The geometry of the device 10 maximizes the apparent production in the restricted area 12 consisting of one or more pins 20 of width b, as shown in Fig. 12b, where the area 12 allows to enhance or change the design bending rigidity of the cantilever 16 , Does not limit the cantilever 16 depends on its entire length l and width w liquid damping. It can also be known that the cantilever 16 will be provided with a conventional electrode (not shown), so that a conventional external measuring circuit (not shown) that provides a bias current can measure the piezoresistance change when the pin 20 is bent. In addition, depending on usage and design choices, external driving force is applied or not applied to the cantilever 16 in a conventional manner.

In the preferred embodiment, there are two pins 20. Assuming that the biofunctionalized end 14 of the cantilever 16 can tolerate a temperature increase of the order of 1k, the length of the cantilever 16 is l, the width is w, the thickness is t, the resonance frequency in vacuum is ω0/2π, and the force constant is K.

We treat this problem as a one-dimensional problem. For the shrinking area, the length is l1 and the cross-sectional area is A, and the beam 16 thermally recessed at the support end 17 supports the substrate 18. In the illustrated embodiment, the cantilever 16 is made of silicon and is assumed to be immersed in water, but any material that can be processed to nanometers can be used, and any peripheral fluid can be considered. The distance x from the connection point of the cantilever 16 and the bracket 17 to the end 14 thereof is measured. For x>l1, the heat loss of the water or liquid into which the device 10 is immersed can be roughly estimated by the relation κSiAdtTdx2=κH2OPT]]>, where P is the circumference of the cross-sectional area A surrounding the beam 16. It is estimated that nT~T/w and d2Tdx2~2(w+t)κH2OκSitw2,]]>where κsi=1.48×102W/mK is the thermal conductivity of silicon, κH2O=0.607W/mK] ]> is the thermal conductivity of water.

When the loss region x<l1, get κSitbd2Tdx2~I2R+4(b+t)TbκH2O.]]> as the boundary condition, and get that the temperature at l1 is continuous, as the heat flux; and for x>l1, the temperature must decrease monotonically.

We consider three example devices 10 of Table 1. For the first type of cantilever 16, a simple thermal conductivity calculation shows that with a steady-state bias current I=250μA, a 1K temperature rise can be obtained at the biofunctionalized end 14, resulting in a power consumption of about 10,670μW. The maximum temperature rise of 12K occurs in the shrinking area 18 about 2.3 μm from the edge of the substrate 18. For this bias current, the responsivity R=/G produced by the first device is about 8 μV/nm.

For the second cantilever 16 in Table 1, the same 12K maximum temperature rise in the contraction area 12 is allowed, which is consistent with the 0.04K temperature rise at the end 14 and occurs at a current of 75 μA. For the device 10, the desired gauge factor is G= 5.2×109Ω/m. Therefore, the desired response rate is 390 μV/nm.

Finally, for the third cantilever 16 in the bar 1, using a maximum temperature rise of 12K in the contraction area 12 and 0.04K at the end 14, a current of 22 μA is allowed. For this device, G=5.3×1010Ω is expected, and therefore, the desired response rate is 1.2mV/nm.

Table 1

For applications driven by liquid coupled thermomechanical noise in the liquid in which the device 10 is immersed, the proportion of coupled fluids in BioNEMS is preferably maximized relative to the background Johnson noise in the device 10. The size of the cantilever 16 determines the relative strength of these two thermal noise sources. These dimensions become part of the amplitude of the liquid damping that determines the intensity of the liquid-coupled thermomechanical noise spectrum in the field of mechanics, and part of the response function of the cantilever 16 (through damping, effective mass, gauge factor, spring constant, and the same for the end Allowable current for heating capacity).

For a fixed thickness t, the signal strength can be greatly improved by reducing the width of the cantilever pin 20 as shown in the curve in FIG. 1. In FIG. 1, for the device with b = 0.4 μm and b = 0.1 μm, compared with the thermal The frequencies of mechanical noise and Johnson noise plot the measured signal strength of the device 10 in units of nV/Hz1/2. Increasing the total length l and the width w of the cantilever 16 will increase the coupling of the liquid, increase the liquid damping and cooling efficiency, and also increase the signal-to-noise ratio.

The curve in FIG. 2 shows the effect of increasing the length l of the cantilever 16 with different lengths of b = 0.1 μm, t = 130 nm and w = 5 μm shown in each example table. The outer liquid is diethylene glycol. Figure 2 is a graph of liquid damping thermomechanical noise, which is predicted to increase relative to the background Johnson noise as the length of the cantilever increases. For a length of 35 μm, the liquid damping thermomechanical noise has a peak magnitude larger than the background Johnson noise. For each cantilever length, select the bias current so that the expected temperature rise at the end 14 does not exceed 1°C, and the maximum temperature in the pin 20 does not exceed 50°C. Table 2 summarizes the bias current used and the estimated temperature rise.

Table 2

In addition to depending on the cantilever size, it should also be noted that the magnitude of the liquid coupled thermomechanical noise depends on the liquid viscosity. Figure 3 shows that for the cantilever 16 with b=0.6μm, t=130nm, w=2.5μm, l=15μm and l1=0.6μm, and with different viscosity ranges corresponding to water, diethylene glycol, glycerol and ethylene glycol The liquid bias is 250 μA, which is the expected signal in the voltage domain for Johnson noise and thermomechanical noise.

Enhancing the probability of receptor-ligand reaction In one embodiment, the device 10 is biofunctionalized by attaching the bioreceptor molecule 24 to or near the end 14 of the cantilever 16, as shown in FIG. 4. Since a basic feature of the NEMS device 10 is the use of bioreceptor molecules 24 attached to the "functionalized" area of the elastic cantilever 16, it is important that for the target ligand 22 of interest, the receptor-ligand binding reaction has the highest probability Coefficient (response probability). However, it is well known that the biologically expressed receptor molecule 24 generally does not have the highest possible binding rate for its target ligand, because there is no selective evolution for this feature. Therefore, it is useful to examine various methods for enhancing the binding rate coefficient of the receptor 24 in the application of the NEMS device 10.

One possible method is to find specific molecules that reduce receptor-ligand binding activity when tightly pedicled with specific types of receptors 24, and use equivalent catalytic reactions. Due to the probability coefficient, the reaction probability is very sensitive to the active energy, for example, it has an exponential dependence, so even if the energy is reduced by a relatively small amount, the binding probability coefficient will be greatly increased. Thus, the gist here is to first attach a layer of such "catalyst" molecules 26 to the small functionalized area 28 of the cantilever 16, and then attach the receptor 24 dedicated to the target ligand 22 of interest. Figure 4 shows the basic concept of this method. According to well-known principles, the ligand-acceptor pair determines the catalyst selected in any given situation.

The second feasible method is to design the receptor molecule 24 with the largest binding rate coefficient relative to the specific ligand 22 of interest. The term "design" here is understood to mean that the use of forced evolution and biochemical techniques is used to select the gene representing the desired receptor 24. When the gene is "isolated" through multiple cycles, the receptor 24 is compatible with the selected Base 22 has a higher binding affinity. This is usually done by inserting multiple copies of the gene of the receptor of interest 24 into the genome of a particular bacterium so that the bacterium will exhibit this receptor on its cell surface.

Then the first generation of bacteria is tested for binding affinity, and those with the highest binding affinity are reserved for the next round of "evolution" cycle. Before starting the next cycle, the change of the receptor gene is generated by gene point mutation, or the change of the receptor gene is more effectively generated by "DNA-gradual movement". These cycles are then repeated until it is clear that the binding affinity cannot increase further. One may hope that by using this technique, the magnitude of the specific receptor-ligand binding rate coefficient can be maximized by at least one order of magnitude.

Non-specific ligand binding effect The curve shown in FIG. 5 represents the possible effects of "background" non-specific binding in response to the NEMS cantilever 16. The curve 30 represents the partial occupancy rate of the functionalized position 28 to the target ligand 22, and the curve 32 represents the partial occupancy rate of the target 22 relative to the background ligand molecule 32. The conditions used are: (1) 1000 target ligand molecules 22; (2) 100 receptor molecules 24; and (3) 10,000 background ligand molecules 32; non-specific binding affinity is greater than the binding affinity of target ligand molecules 22 The combined force is 200 times smaller. The non-specific binding competition effect of the receptor 24 is important for the device 10 shown in FIG. 5.

The signal generation using multiple cantilevers in BioNEMS is in the co-pending patent application entitled A METHOD AND APPARATUS FOR PROVIDING SIGNAL ANALYSIS OF ABIONEMS RESONATOR (serial number (CIT. The situation of a single unactuated cantilever 16 for biomolecule detection has been analyzed in detail. We analyzed the expected detection properties for several driven dual cantilever systems. In the examples described here, we have investigated several general issues related to the use of multiple cantilevers in the BioNEMS device 10. This analysis is discussed from the point of view of signal "design"; that is, we are looking for a device structure that will produce the best signal for the detection properties. The signal is given by the following formula: r(t)=Acos(ω0t+θ)+n(t) 1.1 where A is the amplitude of the signal vibrating at frequency ω0; n(t) is the zero-average Gaussian noise processing with variance σ2n. At this time, if A and ω0 are determined, that is, no fluctuations occur, then the best known detector is a so-called "phase detector" or "lock-in amplifier". The block diagram is shown in FIG. 6. In communication theory, this detector is called a correlation receiver. Therefore, we will examine the device structure that allows us to inject a "carrier" signal with a frequency of ω0, so that the target ligand binding activity will "modulate" the wavelet, that is, change the value of A. The correlation detector of FIG. 6 includes a narrowband filter 34, and the narrowband filter 34 obtains the input r(t) from the device 10. The mixer 38 mixes the reference signal from the oscillator 36 with the output of the filter 34. The mixed filtered signal is input into the low-pass filter 40 and then coupled to the threshold and judgment circuit 42. The threshold and judgment circuit first determines whether the signal can be used as a valid or information-carrying signal, and then if it is valid, the judgment algorithm is executed. It is determined whether the device 10 detects the ligand-receptor interaction of interest. These circuits can be implemented in an analog or digital signal processor or computer controlled by hardware design, firmware or software designed by conventional design options.

However, we have to deal with quite strict constraints, resulting in a huge loss of performance, and the signal amplitude A or phase θ will have significant random fluctuations. This performance loss is shown in Fig. 7, where case A is the case of "no random fluctuations"; case B is the case where there are random amplitude fluctuations; case C is the case where the phase information is lost through random fluctuations. The situation labeled "chopper" is the performance of the signal when the undriven arm 16 is randomly detected using the sum of multiple samples as disclosed in the co-pending application cited above. This method injects a carrier signal and "couples" it with the ligand-acceptor binding activity, thereby serving as a material that is considered to be available for deliberate measurement.

Figures 8a-8c show two possible implementations for generating signals from Equation 1. In Fig. 8a, carrier injection is achieved by mechanically driving the unfunctionalized spiral arm 16a without receptors at a fixed frequency ω0 and amplitude. The functionalized spiral arm 16b responds to the driver 16a through the hydrodynamic coupling of the liquid in which the device 10 is immersed. The driven arm 16b has a piezoresistive portion 46. Both the swing arms 16 a and 16 b are connected to the support 48, and the support 48 can be connected to the base plate 44 or 50.

In Figures 8a-8c, the "no signal" situation is the situation where the spiral arm 16b is "nailed" to the substrate 44 without significant movement; the existence of the "free" target ligand 22 leads to the destruction of this by competitive binding. One condition. Note that the juxtaposition of Fig. 8a is not a preferred structure. In reality, it is best to have two opposite, end-to-end spiral arms. Figure 8c is a slightly different variant of Figures 8a and 8b, in which the no-signal state has two spiral arms 16 and 34 mechanically coupled by a ligand-receptor binding structure. The existence of free target ligands once again destroys this state through competitive combination, resulting in loss of coherent signals.

Although we have discussed multiple spiral arms 16a and 16b, it should be noted that using a single spiral arm 16 with a small area 28 coated with a magnetic film can achieve very effective carrier injection; then an external carrier signal source 36 generates a drive signal, And it is magnetically coupled with the spiral arm 16 to obtain a "constant" amplitude, and the spiral arm beam 16 vibrates in a constant phase. In fact, compared with any other direction, this method can better meet the parameter fluctuation limit. Coupling can also be extended to electrostatic coupling using dipoles or paraelectric films.

The unused use of the multi-rotation arm 16 is to use an array of N identical undriven arms in the average mode or the coincident mode. In the average mode, simply use N outputs to obtain a N improvement when estimating the variance. In the coincidence mode, the restriction condition that a part of the N outputs must be above the threshold during the fixed time of the "signal present" activity is used. In both cases, the method of using multiple spiral arms seeks to increase the ligand capture rate by increasing the number of available acceptors 24.

Reducing the relative fluctuations caused by the fluid Although the single spiral arm 16 in liquid has been well studied, little attention has been paid to the liquid coupling of the spiral arm 16 array. The fluid disturbance generated by the vibrating arm 16 has a long range, and the fluid disturbance is reduced only when the distance between the separating arms 16 is opposite to the power. Moving one spiral arm 16 will generate the motion of the other spiral arm 16 through viscous drag or fluid coupling between the two cantilever arms. This is equivalent to the correlation of the random motion of the spiral arm 16, because the random force generated by the collision of the molecules on one spiral arm 16 will cause the movement of the second spiral arm 16 through this coupling.

This relationship is quantified by the wave dissipation theory. The relative fluctuations caused by the fluid tend to make the correlation caused by the molecular binding between the spiral arms 16 less obvious, and it is more difficult to detect and characterize biomolecules using this mechanism. It is therefore important to understand the correlations caused by fluids in the absence of biomolecules, and to design geometric structures and protocols to minimize them.

Recent experiments conducted by Meiners and Quake on laser-embedded balls provided the first indication of the desired structure. See JC Meiners et al., Direct Measurement of Hydrodynamic Cross Correlations Between Two Particle In An External Potential (Phys. Rev. Lett. 82, 2211 (1999)). The two authors studied the correlation caused by the fluid interaction of two 1 μm rubber beads separated by approximately 3-10 μm. They found that for the closest interval (3 diameter) studied, the inverse correlation of the sphere with the largest inverse correlation is close to the mean square displacement of a single sphere. This indicates that it is difficult to eliminate strong fluid coupling. However, the flow around the long cantilever 16 has very different properties from the flow around the sphere, and a strategy for minimizing the flow dependence is proposed.

The low Reynolds number flow around the moving sphere is in phase with the movement of the sphere in all respects, and decreases by 1/r. The simplified model of the cantilever 16 is to approximate it as a cylinder with a longer radius than that. In this case, the Stokes result can be used for the flow around an infinite cylinder. The flow around the vibrating cylinder is more complicated than the sphere. In fact, for small Reynolds number values, there is no solution independent of Reynolds number, and there is an extraordinary phase relationship between fluid velocity and cylinder velocity depending on frequency and distance. Change the distance scale to aR-1/ Several times of 2. Here a is the cylindrical radius, we set it as the width of the cantilever, and R=ωa2/4v is the Reynolds number, ω is the frequency, and v is the dynamic viscosity of the fluid.

For the BioNEMS cantilever 16, R is usually about 1. This velocity field is the cause of the movement of the second cylinder, so the movement of the second cylinder caused by the force on the first cylinder is dependent on the distance and frequency. By specifying that the fluctuations in the system are proportional to the damping in the system, the fluctuation dissipation theory gives the dependence of distance and frequency on the correlation of noise. The velocity field curve with the distance r/a from the cylinder axis as a function in Figure 9 shows that in fact, different integrals of the velocity field have zero points at different distances, indicating that in one or other phases of the liquid caused by related noise , You can find the parameters corresponding to the zero point (cylindrical spacing/radius and frequency). 9 shows the fluid velocity component of the velocity of the vibrating cantilever 16 parallel to the velocity of the vibrating cantilever 16 as a function of the distance r in units of the cantilever width a for the Reynolds number R=1. Note the extraordinary phase relationship given by the integral components of the real part (in phase) and the imaginary part, as well as the zero points of the different integral components.

The calculation of the correlation allows the optimal process to be designed as follows. The cross-correlation between the displacements of the two cantilevers C12=<x1(0)x2(t)> and the precise relationship between the deterministic fluid coupling between the cantilevers are as follows: ddtC12(t)=-kBTX12(t)--- -t>0----4.1]]> where X12 is the "magnetic susceptibility", which gives the force F1 acting on the first cantilever 16 and the displacement x2(t) of the second cantilever 16.

<x2(t)>=-X12(tt)F1(t)dt---4.2]]>Angle brackets indicate the overall mean, again emphasizing certain estimates, calculations or time energy Quantify random motion. The magnetic susceptibility X12 can be calculated from the Stoke velocity field depicted in Figure 9, and the noise correlation can be predicted, but it is not shown here.

Incorporating NEMS into Microfluidics FIG. 10 shows a schematic cross-sectional view of the device 10 in which the cantilever 16 is coupled to the microfluidic flow channel 52. The area etched through the wafer 54 forms part of the final liquid flow path. The entire array of cantilevers 16 can be fabricated within a single flow channel 54. There may also be multiple channels 54 and different devices 10 are in different channels 54 depending on the desired application. The liquid flow channel 54 can also be directly surface-fabricated in silicon. Figures 11a-11c schematically show the array 56 of the device, which has three perspective views of increasing scale, showing a plurality of cantilever arms 16 supported on parallel supports 48, in communication with the inlet and outlet liquid flow channels 54 16 rows of parallel cantilevers are formed in a single liquid flow channel 58.

Manufacturing of BioNEMS Thin-Film-Based Device The flowchart shown in FIGS. 13a-13g schematically shows the steps involved in the manufacturing of the thin-film-based BioNEMS device 10. Starting with the side cross-sectional view of FIG. 13a, the fabrication of these devices is started by placing a silicon device layer 148 on insulating wafers 152, 154, for example, a 375 nm SiO2 layer 152 on a 675 μm Si layer 154. The buried oxide layer 152 must be thick enough to serve as a stop layer through the subsequent etching step of the wafer 154. For most devices under consideration, the Si device layer 148 should have the required thickness, and for the undoped portion of the silicon cantilever, the thickness is between 20 nm and 100 nm. In the illustrated embodiment, an 80 nm Si layer 148 is provided under the 30 nm heavily doped Si layer 150. The resistivity of the layer 148 should be high relative to the heavily doped layer 150 to be grown (10 Ωcm is sufficient).

The back side of the wafer 154 is polished. This is necessary to bond the elastic material there, where the microfluidic channel 52 will be defined as described below. At the same time, the wafer 154 can be thinned to reduce the final volume of the liquid flow channel 52 and the thickness that must be etched in the subsequent etching step. The final wafer thickness of 300 μm is reasonable for maintaining the structural integrity of the wafer 154 while reducing unnecessary materials.

Then, a heavily boron-doped silicon layer 150 is epitaxially grown on the top surface of the layer 148, and the layer 150 will form a piezoresistive conductive layer forming part of the NEMS device 10. For most devices 10, this layer 150 is between 7 nm and 30 nm thick. The resistivity must be lower than the underlying layer 148. Usually the doping level is 4×1019cm-3.

The next step involves etching through the back side of the wafer 154 to create a thin film, as shown in the side cross-sectional view of Figure 13b. Bosch deep reactive ion etching (DRIE) can be used to form trenches 158 through layer 154. For this step, as shown in Figure 13b, it is sufficient to use a photoresist or oxide mask 156 of approximately 6 microns. A 50μm2 film is used, but it is arbitrary, and the size used depends on the application. Then, hydrofluoric acid is used to remove the oxide layer 152 just below the silicon film 148 from the bottom of the groove 158, as shown in FIG. 13c, to define the area that will become the film 162.

Then, on the top side of the device 10, lithography and metal deposition are performed in alignment with the film 162 on the layer 150 to form a contact pad 160, as shown in the top view of FIG. 13d, which represents multiple dice formed at the same time. For the bonding pad 160, 30 nm of chromium (as an adhesive layer) and then 250 nm of Au are deposited into a desired pattern to form an ohmic contact with the boron-doped silicon layer 150.

Then a silicon nitride layer 174 (300 nm) is deposited on the Au layer, and then a 200 nm silicon dioxide layer 175 is deposited to passivate the device 10, as shown in the top view of FIG. 13e, and then a chromium layer 176 is deposited, as shown in FIG. 13f It is shown in the top view to protect the silicon dioxide 175 during the manufacturing process and to provide electrical connections at the height of the step formed by the passivation layer, as shown in FIG. 13g.

Then using electron beam lithography, first pattern the gold solder joints (not shown in the figure) at the end 14 of the cantilever 16 to be biofunctionalized, then pattern the cantilever 16 (on PMMA), and then evaporate the 30nm chromium layer 178 (Not shown in the figure), and peel off as shown in the perspective view of FIG. 13i. This part of the manufacturing process consists of two steps. The first step is to deposit a gold square (not shown in the figure) at the end 14 that will become the cantilever 16, which is used for biofunctionalization together with an alignment mark (not shown in the figure). The second step is a photolithography step, which masks the area including the cantilever 16 with a chromium layer 178 (not shown in the figure).

The device 10 is suspended using vertical plasma etching (NF3, Cl2, Ar), which removes the unshielded portions of the thin film defined in the layers 150 and 148, resulting in the cantilever 16 shown in the top view of FIG. 13i and FIG. 13j. The chromium mask 178 (not shown in the figure) is then removed using wet etching, and the sample is dried with a critical point dryer. The wafer is cut into small pieces as shown in the top view of Figure 13k.

Then using patterned photoresist as a model, the microfluidic channel 52 is made of silicone elastomer, and then the photoresist is etched to define the actual liquid flow channel in the molded elastomer capsule 182, as shown in Figure 131 Show. When placed in an oven at 85°C for 24 hours, the silicon of the device 10 self-seals the flow channel 52. Then, as shown in FIG. 13m, the device 10 or the gold solder joint 180 is biofunctionalized by a conventional method, such as flowing a liquid through the flow channel 52, which carries acceptor molecules that preferentially bind to the gold solder joint 180.

The damping produced by the pom-poms is to provide binding receptor molecules 62 only for the pom-poms 60 of large dissipative molecules. The pompons 60 and the binding molecules 62 float freely in the liquid. The binding molecule 62 is suitable for binding to the ligand of interest. The end 14 is also biofunctionalized using receptors suitable for binding to ligands. The trapping of the receptor on the end 14 of the ligand of interest to which the receptor 62 and the pompons 60 are attached will cause the damping coefficient of the end 14 to increase sharply.

FIG. 14 is a mathematical model of a cantilever 16 with a mass of M. The cantilever 16 has a pom-pom 60 with a molecule 62 as a spring and attached to the end 14 thereof. The pompom 60 is designed to maximize the dissipation and therefore the noise, and is composed of star-shaped dendrites. Molecule 62 may include an alkane or ligand chain. For the system in Figure 14, the formula is: Mx··+γx·+κx=km(xd-x)+F]]>γdx·d=-km(xd-x)]]>xd =kmxiωγd+km]]> where M is the mass of the cantilever 16, γ is the fluid damping coefficient of the cantilever 16, κ is the spring constant of the cantilever 16, x is the displacement of the cantilever end 14, km is the effective of the molecule 62 The spring constant, xd is the displacement of the pompom 60, γd is the fluid damping coefficient of the pompom 60, and F is the external force applied to the cantilever 16.

The equation of motion for the system can be rewritten to indicate the effective damping and effective spring constant of the system given by the following formula: γ~=γ+γd1+(γdkmω)2]]>

κ~=κ+κm(γdωkm)21+(γdωkm)2]]>Choose pompom 60 to make its damping coefficient as large as possible to maximize dissipation , So as to maximize the noise, so: γd=km/ωγ~=γ+12kmω]]>The relative loss of biomolecules and pompoms on dissipation and noise corresponds to the following magnitude: Δγ&gamma ;kmγω]]> Without departing from the spirit and scope of the present invention, those of ordinary skill in the art can make various changes and modifications. Thus, it must be understood that the illustrated embodiments are given for illustrative purposes only, and should not be seen as limiting the invention defined by the following claims. For example, although the elements of the claims are given below in a certain combination, it must be understood that the present invention includes other combinations of fewer, more or different elements. Even if the right to these combinations is not initially claimed, these combinations are also disclosed above. .

The words used in this description to describe the present invention and its various embodiments should not only be understood in the meaning of its usual limited meaning, but also include the scope of the meaning outside the scope of the ordinary definition, the structure, material, or function of this description definition. Therefore, if an element in this description is understood to include more than one meaning, its use in the claims must be understood as common to all possible meanings supported by this description and the word itself.

Therefore, the definitions of the words or elements in the following claims are defined in this description to include not only combinations of elements given literally, but also all equivalents that perform substantially the same functions in substantially the same manner to obtain substantially the same effects. Structure, material or function. Therefore, in this sense, two or more elements can be equivalently substituted for any one element in the following claims, or a single element can be substituted for two or more elements in the claims. Although the elements are described above as acting through a certain combination and are initially required as well, it is clearly understood that in some cases one or more elements of the claimed combination can be removed from the combination, so The claimed combination may involve a sub-combination or a variant of the sub-combination.

It is believed that insubstantial changes to the required subject that are known or conceived in the future by those of ordinary skill in the art are equivalent within the scope of the claims. Therefore, obvious alternatives that are known to those of ordinary skill in the art now or in the future are limited to being within the scope of the defined elements.

Therefore, the claims should be understood to include the content of the above specific description and description, obviously alternative content, and the content that essentially contains the basic idea of the present invention.